Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 6 SOCIAL AND ECONOMIC CONSEQUENCES OF TEENAGE CHILDBEARING Sandra L. Hofferth I=RODUCT ION The assumption of active parenting signif icantly changes a young woman's or man' s life. As abscessed in previous chapters, caring for an infant takes time and energy which is therefore not available for other activities. Although parents are never fully prepared, those who are mar r led, with a reg ularly employed wage earner and a reason- ably stable existence generally have the resou roes to cope adequately . The demands of parenthood must cone as a shock to the unmarried teen- ager who is enrolled in school, who is dependent on her parents, and who knows very little about caring for children. The first part of this chapter focuses on the long term conse- quences of ear ly childbear ing for the mother, the father, and other family members. The major objective is to compare some ten years after high school the economic s ituation of young women and men who bore (fathered) a child as a teen with that of others who delayed child- bearing until their twenties. The questions that will be addressed are the following: 1. Are there effects of early childbear ing on the later social and economic well-being of the mother, the father, and other family members net of initial differences between early and later childbearers? 2 . I f there are effects, how do they operate? That is, through what mechanisms or intervening f actors do they operate? 3. Have these effects changed over time such that early chilabear- ing has more (or less) ser ious consequences for recent birth cohorts of young women and men than for earlier bi rth cohorts of young women and men? Research has shown substantial car iation among early childbearers in economic well-being, and it is important to know why some do well and others don't. Thus an additional question will be addressed: 123
124 4. Among early childbearers themselves, what factors differen- t late those who are doing well f ram those who are not doing well? Part Two focuses on the consequences of early childbearing for society. Finally, Part Three focuses on the hypothetical impact of policy interventions. The perspective used in this chapter is that of the life course, The social patterns in the timing, duration, spacing and order of events. (Elder, 1978:211. One of the central features is the notion of Multiple interdependent pathways (career lines) from birth to death. (Elder, 1978:221. Such career lines occur in the marital, parental, and socioeconomic spheres. The relationship between the timing of events in these different spheres represents an important character istic of individuals. There are also regular patterns across individuals. For example, a majority follow a coon pattern regard- ing timing of school leaving, entry into employment, marriage and childbearing ~ see for example, Hogan, 1980) . Myth multiple career lines, the scheduling of events and obligations becomes a basic problem in the ~nanagement of resources and pressures. {Elder, 1978: 27) . Parenthood is an event that radically affects the life of the mother. The demands of a child simply cannot be ignored without risk. Thus the timing of parenthood relative to other career lines is a major concern. In this chapter we will consider schooling, marriage, and employment as other interdependent career lines and explore the inter- relationships among events in these different domains. The ultimate test of the importance of timing and sequencing of events is the echo comic circumstances of the individual at some later point in life, in particular, own income, income of other family members, poverty status and welfare dependence. Direct Versus Indirect and Total Effects Just because research identifies no direct causal connection be- tween two var tables, for example, between the age at which a woman has her first birth and family income, for example, does not mean there is no association at all. For example, if an early first birth is ascot c fated with reduced schooling, which is, in turn, associated with lower earnings, and lower income, then an early f irst birth is indirectly associated with lower family income later on. The total of the direct impacts of age at first birth and its indirect impacts through other variables is called the total effect of age at f irst birth. The path- ways through which a variable such as age at first birth affects ~rari- ables later in life explain the impage of age at first birth. That is, they explain how it can affect later well-being without there being any direct causal connection.
125 Measurement of Early Childbear ing Mos t of the stud ies ref er red to in this chapter measure the age at which the young woman or man bore ~ fathered) a f irst child in single years of age. This is then associated in a statistical analysis with education, number of children or family income. The coeff icients re- ported, therefore, show what difference there is in years of schooling, for example, between youth who differ by one year of age at first birth. The assumption of the model is that the effect of delaying a birth for one year is the same whether a young man or woman delays from 16 to 17, f ram age 20 to 21, or f ram age 26 to 27. This is a strong assumption, and one that may not necessarily be true. As an alternate t ive, then, some of the models looked only at a subsample of teenagers. In this case, using the same age measure, the results indicate the difference that delaying a birth for one year during the teens years makes in the outcome measure. This may be more useful in policy terms, but it then does not compare teenage with older childbearers. Another way to compare the effects of teen versus older child- bearers would be to simply a ichotomize at age 19, for example, and com- pare the socioeconomic status of those with a f irst birth at or before ace 19 and those with a f irst birth after age 19. The choice of the cutoff point then becomes an issue, since it may greatly affect the results. None of the studies cited here dichotomized the age at f irst bi rth var table. Given the fact that the models included here are linear models, in addition, the types of relationships between age at f irst birth and soc ioeconomic outcomes are severely constrained. The reader therefore is cautioned that the research reported here, while of very high q~al- ity, is limited in its sensitivity to complex relationships. CONSEQUENCES FOR THE MOTHER, FATHER AND OTHER FAMILY MEMBERS The f irst question is how, once they have reached their late twen- ties and early thirties, do women and men who had their first child bee fore twenty and those who had it at ter age twenty compare on economic well-being? Which events and domains account for most of the relation- ship we f ind? Second, among early childbearers not all are doing poorly. What determines differential adaptation? Some potential ex- planatory factors include a) individual differences in background, aspirations, motivation and ability, b) resources: family socioeconomic status, informal support networks; c) formal programs of social inter- vention, and d) career contingencies: other events occurring around the time of the birth in other career lines e.g., marriage, e~nploy- ment, schooling. Most of this research focuses on young women; rele- vant data for young men are presented where available.
126 School ing The most general sequence of schooling, marriage and childbearing is that of completion of schooling, marriage, and then childbearing. In this section we will focus particularly on the ti~ning/sequencing pattern in which childbearing precedes school completion. Because of the time and energy that raising children require, which interferes with the time and energy required to study and attend classes, women who bear a child dur ing the school years of ten leave before they can complete their schooling. This is especially the case for those who bear a child during the high school years. Results from a number of studies show that young women who bear a child as teenagers are subs stantially less likely to complete high school than those who bear a child later on. All the studies reviewed show that early childbearers exhibit a substantial educational def icit relative to later child- bearers. However, studies have also found substantial preexisting differ- ences between early and later childbearers, differences that may ex- plain the difference in completed schooling. Card ant] Wise (1978) for example, found that young women who bear a child while in high school not only were of lower sac ioeconomic status when they were in ninth g race, but already had lower academic abilities and lower educa- tional expectations than their classmates, factors which also predict poor school performance and poor later life chances. With the except tion of one study (Rindfuss et al., 1980 I, every study that has been able to control for initial differences between early and later child- bearers (Card and Wise, 1978; Haggstrom et al., 1981, 1983; Roo and B ilsborrow, 1979; Hof forth and Moore, 1979; Mar ini, 1984 ~ has found an additional impact of having an early birth. Thus the bulk of the evi- dence is that there is an additional impact on school completion of having a ch ild at an early age above and beyond the impact of the initial disadvantaged situate of those who tend to have births at an early age. The impact of an early birth has also been shown to be greatest during the high school years tHofferth and Moore, 19791. This does not negate the fact that some young women do drop out of school physically or even mentally far before bearing their first child. There in evidence that a sizeable proportion (one~quarter to one~third} dropped out prior to a first pregnancy (Morrison, 1984~. However, even among those with poor school records, those who have a first birth while in high school face even greater odds against cow plating their schooling than those who delay that f irst birth for several years. Sane attention has been given to the issue of whether the relation- ship between - Cooling and dropping out of school has changed over time. The C- and Wise study looked at the earliest birth cohorts-- born in 1942~ and 1945-46. The Hofferth and Moore study looked at birth cohorts ~ ~44-54, while the Haggstrom et al. study looked at a cohort born in l9S4, approximately. It is possible that some of the differences between the results are due to changes over time. Mot t and
127 Maxwell, for example, found that young women were more likely to stay in school following a first birth in 1979 than they were in the late 1960's. However, this does not reduce the disadvantage they suffer. McCarthy and Radish (1982) show Chat even though early childbearers complete more schooling than they used to, their childless peers are also completing more schoolgirl. Thus they are not better off in rela- tion to their peers. Since so much reliance is placed on schooling today, they may be even worse off. One study suggests that a small additional amount of schooling for these young women does not boost their earning power enough to keep them f ram needing public "s istance {Moore and Wertheimer, 19827. How much schooling and what type is needed to mace a signif ican~c dif ference in their economic well-being would appear to be valuable questions to answer. There are several other issues here. First, what are the factors med dating whether an adolescent childbearer remains in school or not? One of these is the legal system. Until the mid 1970s, young pregnant women were of ten not permitted to remain in school. Title IX of the Education Amendments of 1972, which was implemented in 1975, prohibits discrimination because of pregnancy or parenting status in publicly supported educational programs. Schools make a variety of arrange- ments for the schooling of pregnant students, frown keeping them in regular classes to providing separate programs ~ Bellman, 1982} . A1- ttlough these efforts vary in quality, they appear to have had a sub- stant ial impact on school couplet ion (Hott and Maxwell, 1981} . How- eve r, even so, car ing for a young baby puts an enormous burden on a young women. What other factors have been shown to be associated with keeping a pregnant adolescent in school? Family support has been shown to be important to whether or not an adolescent childbearer remains in school (Furstenberg and Crawford, 1978~. Those who do not marry and who remain at home with their parents are more likely to complete high school than either those who do not marry but move way from home or those who marry. Enrollment in spec ial school prog rams may also af feet school come pletion. In their 17 year follow-up of adolescent childbearers, Furstenberg and Brooks~Gunn (1985) found a strong association between staying in school and attending a special school for teen mothers. S. ince adolescents who had higher ambitions were much more likely to pa ~ t ic mate in the spec ial p tog ran than to stay in the regular school, and more highly motivated adolescents did better later on regardless of the type of school, this may explain their differential school con- tinuation. However, after controlling for its selective attraction to mot ivated adolescents, Furstenberg found that the assoc iation between attending the school and later well-being remained strong (Furstenberg and Brooks~Gunn, 19851. Those in the special school for pregnant girls did substantially better in later life than those attending a regular school program ( and those who dropped out) .
128 An important issue, one which has not received much research atten- tion is that of identifying factors associated with whether or not a young woman who has had an early birth and has dropped out of school returns to complete her schooling. Research evide~x:e suggests that at least during the first decade after a birth, early childbearers (who have dropped out) do not return to school at a higher rate than those also out of school but who have not had a f irst birth (Hove et al., 1978:29-30). Thus they are not likely to catch up. Research comparing early childbearers and delayers at a later point in the life cycle shows that although a substantial proportion of pregnant adolescents do d rop ou I, a substant ial p roport ion do eventually return to complete additional schooling or receive a GED (Furstenberg and Brooks~Gunn, 1985; Mott and Marsiglio, 1985~. Unfortunately, the evidence shows that a year of school attendance is not assoc iated with completing an additional year of education (Furstenberg and Brooks~Gunn, 19851. Thus, al~chough early childbearers do return to school, it takes a lot longer for them to complete a year of schooling than it does for those who d id not d rap out . In add it ion a GED may not be as advantageous as a high school diploma. Given initial differences and the cumulation of disadvantage, it seems unlikely that early childbearers will return to school at higher enough rates after their children are q r own than later childbearers to reduce their relative disadvantage. The data show a declining difference in educational attainment with age, but one which remains substantial and which does not disappear (Card and wise, 1918~. If, in fact, as has been suggested by other research (Card, 1981; Newcomer and Udry, 1984), their daughters bear children at ear ly ages ~coo, these mothers may continue to have childrear ing dut ies for many more years. Although most research has focused on females, there is reason to believe that fathering a child may also have consequences for males. Are men who father a child at an early age more likely to drop out of school? If so, do they eventually receive accreditation in the form of a GED? Is ever father ing a child associated with less schooling or are any effects limited to those who live with their children? F inally, are the differences due to substantial preexisting differences or to the early birth itself. Card and Wise (1918) showed that half of all women and seven out of ten men who had borne a child before age 18 completed high school by acid .99, compared with almost all who delayed childbearing until their eat: twenties. The differences among early and late fathering males are less striking than among early and later childbearers ~ female), but nonetheless are important. A recent analysis (Marsiglio, 1986} based on the National Longi- tudinal Survey of Youth, waves 1979-1983, also found, net of factors such as parental education, family structure, race, region and reli- gion, that young men who reported fathering a child during their teen years had completed signif icantly fewer years of schooling by 1983 than those who did not report having fathered a child. This research, however, did not control for differential IQ and aspirations among fathers and nonfathers. Thus it cannot be concluded that this effect
129 is due only to the birth of the child. The authors failed to find evi- dence that living with a child has Snore impact on a father than ever having fathered a child. However, given the high degree of instability of living arrangements of young parents, this may not be surprising. It would be helpful to have information on whether The father is, in fact, in touch with the mother of the child and whether the father con- tr ibutes to the support of that child . Parenthood All the evidence supports the conclusion that early childbearers have more children, especially more unwanted children, and that they have ached more rapidly than older childbearers Russell and Menhen, 1978: Furstenberg, 1976; Prosser, 1976; Moore and }lofferth, 1978; Roo and Suchindren, 1979 ; Bu~npass et al., 1978) . The issue of whether this relationsh ip has changed over time is an important one. There is e~ri- dence that the dif ference between the earliest and later childbearers is declining with more recent birth cohorts as a result of greater fertility declines among teen mothers (Millman and Bendershot, 1980~. If this result holds up it will be an important one, since the dif- ference in family size is the largest and clearest difference between early and later childbearers, and, as we shall see in the following pages, has the most implications for later well-being. How can the difference between early and later childbearers in family size be explained? One potential explanation is that early childbearers have a longer period of exposure to childbearing. How- exrer, the difference in family size by age at f irst birth holds even controlling for length of exposure (Trussell and Menken, 19781. A second possible explanation is that the youngest women are the least likely to have used contraception at f irst intercourse and least likely to use it consisten~cly thereafter. This does appear to be supported by research evidence (Zelnik et al., 1981) . A third pos- sible, but untested, explanation is that young women who start their families early are familistic in or ientation and want to have larger families. This could be the case for those who intended the f irst birth: however, this accounts for only a minority of teenage first births--23 percent according to Zelnik and Kantner (1978~. A fourth possible explanation i. that early childbearers are less able to take a future or ientation and to plan. AS a result they have Snore unwanted pregnancies across the life span (C~etkovich, 19807. This hypothesis has not been tested. It is clear that differential schooling also increases the gap between early and later childbearers in f amily size. Research has f ound evidence that young women with more schooling are better con- traceptors, and, therefore, are better able to limit their family size. They also desire fewer children. Thus, the age at which a woman has a first birth indirectly affects family size through the schooling she obtains.
130 What factors differentiate early childbearers who have large and small families? When Furstenberg and Brooks~Gunn (1985) went back and reinterviewed their adolescent mothers after 17 years, they found, in contrast to what they expected, that only a relatively small proportion had gone on to have large families. Most had been able to control their fertility. The method that they used was sterilization; about half of these mothers had been Inter illumed for contraceptive purposes. Thus the fertility of these mothers was comparable deco that of delayed childbearers in other surveys. Those young women who were able to con- trol their fertility and, therefore, had the fewest children at the 17 year follow-up, were those who had been at grade for age, who had had high educational aspirations, who used birth control, who were enrolled in school, who delayed a second birth, and who were not married at the f ive year follow-up. Those who attended a special school and those who were in a special hospital program were more likely to use birth control and, as a result, likely to have a small family 17 years later. Although early childbearers have larger families than later child- bearers, hypotheses reasons for thin association have not been tested. Recent research tHeckman et al., 1985} suggests that differences bee tween early and later childbearers that existed prior to the f irst birth may explain the association. If so then what these differences are need further explora~cion. F inally, no research in the consequences of early childbear ing on family size have been conducted on males. Such analysis depends on reports of births, and males substantially under report such events. The quality of male data needs further study (see Marsiglio, 1986~. Marriage and Marital Dissolution There is a very strong relationship between mar ital and parenthood careers. Although the Host colon sequencing pattern is for marriage to precede pregnancy and birth, premaritalpregnawy, marriage and a postmarital birth has not been uncommon. A pattern of increasing imp portance is that of a birth followed by marriage. There :s a strong relationship for whites between the age at which a woman has her f irst child and her age at first marriage; the relationship is weaker for blacks. Wertheimer and Moore (1982) showed that a birth to a woman aged 15 to 1? increased the probability that she would marry from .07S to .240 if she was white and from .056 to .110 if she was black. Recent data show that 96.5 percent of firstborn black babies to women 15 to 19 were conceived out-of-wedlock in 1980-81, compared with 64.4 percent of first born white babies to women 15 to 19, and 87.9 percent of black mothers and 36.8 percent of white mothers were still single at birth (O'Connell and Rogers, 19841. The proportion who have married with in 2 years is also smaller for blacks and whites. Among those who eventually marry, whites merry much sooner than blacks. According to recent data, 53 percent of whites (who eventually marry) were married in 3 years, compared with 29 percent of blacks. Data also suggest that
131 the longer the period of time between birth and marriage, the less likely the mother is to marry the father of the child (Furstenberg, 19761. Thus young women who have an early first birth are more likely to marry soon thereafter, although this relationship appears to have weakened over the past decade and to be especially weak for blacks. The first question is what is the effect of an early birth on marital disruption, relative to that of an early marriage? There are several possible hypotheses as to the effect of the timing of marriage relative to a birth on disruption. First, the literature to date shows that an early marriage is consistently associated with divorce or sepa- ration (Glick and Norton, 1977; McCarthy and Menken, 1979; Weed, 1974; Bumpass and Sweet, 19727. The intervening mechanism may be the youth- fulness of the partners, their lack of exper fence with other potential partners, and the extent to which they have yet to experience important adult transitions. In contrast, some research finds (Furstenburg, 1976; Card and Wise, 1978; Furstenburg and Broolcs~Gunn, 1985; McCarthy and Menken, 1979) that an early birth increases divorce and separation f or men and women. The mixed evidence may be due to ~ related phenol enon. During the early years of marriage, couples with a young child have a substantially lower probability of divorce relative to child- less couples. The presence of a young child appears to depress divorce, at least during the early years of a marriage. It is hard to disentangle the influences of an early marriage and an early birth, since marriage and childbearing are tied so closely together, aspen cially for whites. The relative influence of early marriage and early birth cannot be tested among whites, for example, because these factors a re so h ighly assoc iated . Howell r, th is hypothes is might better be tested among blacks s ince blacks have a much lower probability of marrying soon after a f irst pregnancy (and after a f irst birth) and are unlikely to marry before pregnancy ~ teenagers) . That is, among black teens, a pregnancy is much less likely to precipitate an early marriage. In fact, it is only among blacks that an early first birth is associated with lacer marital disruption, net of early marriage (Moore and Waite, 1981) . Thus it is still too early to rule out an additional impact of a premarital birth or of a short birth interval. However, it is possible to tease out the differential impact of marriage timing among those who bear a f irst child as teenagers. Young women and men who marry soon after a pregnancy may be better off than those who wait until after the birth; however, they Nay be more likely to divorce achy those who marry later, which Nay make them even less secure economically. Probably the most important question is what is the differential divorce proneness of marriages contracted before pregnancy, after pregnancy but before a birth, and after a f irst birth? Research shows that teenage mothers are less 1 ikely to exper fence a marital separation if they marry before the birth than if they marry after the birth; there is little difference in divorce probability bee tween those who marry before versus after becoming pregnant ~ but before the birth) (McLaughlin et al., 1984) . Differences in divorce probabili- t ies by mart iage timing are relatively short term for blacks, but have
132 longer term effects for whites. The impact of marriage timing appears to be declining over time, as it had no impact on divorce/separation among recent birth cohorts of young women. What are the potential explanations of the differential impact of marriage timing? First, young women and men who marry before the birth may be different from those who don't in ways that affect marital sta- bility. In particular, they may be more committed to their partner, in more stable situations, and so on. The researchers controlled for a variety of background factors that could potentially also be associated with disruption (McLaughlin et al., 19847. Thus the possibility of other differences, while still present because of the limited nature of variables that are available, is minimized. A second possible explanation is differential schooling. The amount of schooling the young woman had attained at marrisqe was not associated with the probability of separation, however (McLaughlin et al., 1984) . One factor that was associated with a higher probability of separation was whether the f irst birth was unwanted or mistimed. An unwanted or mistimed f irst birth was associated with ~ higher proba- bil ity of divorce or separation. Furstenberg and Gunn {1985) found substantial marital instability among their adolescent childbearers. Almost. all eventually married 78 percent. However, about 2/3 of first marriages ended; by 17 years after the first birth only 26 percent of the sample were currently mar- ried in a first marriage. Two fifths were previously married, and 8 percent were currently married in a second or later marrisqe. The authors concluded that adolescent parenthood seriously damages a women's prospects for a stable marital union. What is not known is this relationship holds--whether it is due to the child or to other factors that affect both marital instabilityn and early childbearing. Final'-', very little is known about the characteristics (and prospects) of the men these early childbearers marry or could marry. work Labor Force Participation and Hours Worked The research suggests that the age at which childbear ing beg ins is not as important as the length of time since the (most recent) birth in influencing whether or not a woman works. Having a young child consis- tently lowers labor force participation, whereas an early birth does not. Of the three stud ies that have specif ically addressed this issue, one {boo and Bilsborrow, 1980) finds no effect of early childbearing wh ile two stud ies f ind a weak pos it ive ef feet of early childbear ing on labor force participation (Rofferth et al., 1978; Card, 1919} . In these studies early childbearets {female) appear to be somewhat more likely to be in the labor force 10 years after high school than later childbearers. This is probably due to several factors: 1) Since early
133 childbearers start their families early, at ~ and 5 years after high school fewer early than later childbearers are working (Card, 1977) . Ten years after high school, however, ached children are older while later childbearers have just begun their families and have young chil- dren in the home. THUS the early childbearers were more likely to be working 10 years after high school in the Card study and at age 24 in the Hofferth et al. study. 2) Early childbearers may have a greater economic need to work. Nearer married mothers who had an early birth have a high likelihood of being employed tHaggstrom et al., 19817. In a related study Trussell and Abowd ~ 1979 ~ also found that among whites increasing age at first birth lowers the propensity to work by raising the wage required to attract them into the work force. There are sex differences in the association between early child- bearing and employment. At 1 and S years out of high school more males in the adolescent childbearer group were working, compared to their classmates (Card, 1977) . Thus for finales, each parenthood leads to entrance into the labor force. However, by 11 years out, these differences had disappeared. By 11 years after high school most non- parenting males had also completed their schooling and entered the work force so the duff ference disappears. Females, in contrast, work less while they have young children in the inane, but as their children mature, they return to work. Thus the timing of the birth affects when that hiatus will occur. By the mid twenties, the later childbearers are beg inning their families and drops ping out of the work force while the early childbearers are reentering. Work Exper fence This is the only area in which there is any disagreement among the various studies, and this disagreement is not hard to resolve. Two studies {McLaughlin, 1977, and Koo and Bilsborrow, 1979) found that, controlling for age, education and socioeconomic background, early childbearers accumulated more experience after the birth of the first ch ild (McLaughlin) or after marriage (Koo and Bilsborrow) than later ones. McLaughlin and Koo and Bilsborrow hypothesize that early child- bearers have a greater economic need to work than later ones. In con- trast, Hofferth et al. {1978} show no relationship between age at first bi rth and proportion of years worked since age 18 by age 24, net of other factors. In a study that looks at work experience at age 27, Hofferth and Moore (1979) found that later childbearers actually have accumulated more world experience since age 18. Again, these differ- ences are probably a function of the time period over which experience is measured. The former two studies looked only at experience follow- ing a birth or marriage while the latter looked at experience since age 18. Later childbearers probably worked more than early child bearers pr for to marr iage/birth, while less following marriage/birth. Thus the d if ferences in results between the several studies are ex- plainable. Experience depends on where you start to accumulate it. No comparable data are available for males.
134 Occupational Status Researchers have measured occupational status in a variety of ways: the National Opinion Research Center's Occupational Prestige Scale and the Duncan Socioeconomic Index (SKI) are the most colon. In general, researchers find no direct effect of a woman's age at first birth on later occupational status or prestige. coo and Bilsborrow found no in panic of age at first birth on occupational prestige scores of women 35 to 39 and 40 to 44 in 1973, controlling for a variety of background factors, education and work experience. Using the Duncan Socioeconomic Index (SET) Hofferth et al. (1978) found no different between early and later childbearers at age 24 in occupational prestige. Haggetrom et al. (1981) found that scores on a career index similar to the SKI dif- fered little by birth timing. McLaughlin (1977) used measures of earning potentis1 in the short term and the long term, as he called them. The short term measure was the median 1959 earnings of all women working fulI-time in the first job held at least 6 months within the first five years after the first birth. The long term earning potential was median 1959 earnings of all women working full time in the occupation held currently or most rem century for at least six months. For both measures there was a positive but non-significant direct impact of age at first birth net of educa- t ion, exper fence and sac ioeconomic status. Finally, Boo and Bilsborrow (1979) also failed to find any direct impact of age at first birth on the husband's occupational status. Even though no direct effect of age at first birth on occupational status was found, there do appear to be come indirect effects. Card (1977) found age at first birth to be a determinant of occupational prestige for both men and when 11 years out of high school, net of background factors such as race, SES, aptitude and educational plans held in high school. A stronger relationship was found for women than for men. Other research has shown a strong relationship between educa- tional attainment and occupational status, and between work experience and occupational status (McLaughlin, 19771. To the extent that age at f irst birth reduces schooling completed, it is likely to reduce occupa- tional statue later on. The effects of age at first birth on work ex- perience are somewhat unclear. McLaughlin (1977) concluded. that the strongest indirect effect operates through education. Economic Well-Being Women's Mourly Wages and Annual Earnings The evidence is consistent across all studies: there is no direct impact of early childbearing on women's hourly wages Hofferth et al., 1978; McLaughlin, 1977; Roo and Bilsborrow, 1979; and Trussell and
13S Abowd , 1979 ) . The same appears to hold for males (Haggstrom et a l., 1981) . The evidence consistently f inds no direct impact of the age at first birth on female earnings, net of other factors (Woo and Bils- borrow, 1979; Hofferth et al., 1978; McLaughlin, 1977; Haggstrom et al., 1981~. However, there do appear to be indirect effects. Card ( 1977) found that with only controls for background variables, adoles- cent childbearers earned less than later childbearers or those child- less at all follow-ups. Other researchers have specif fed these inter- vening ef Sects. An early birth inc teases family size, which reduces the proportion of years worked and the hours worked last year, which reduces earnings at age 27. At early birth reduces schooling, which reduces the proport ion of years worked and reduces hours and earnings at age 21. Add ing all the ef feats up, early age at f irst bi rth is associated with reduced earnings, but this is because it is associated with reduced schooling and increased family size. The length of time since (most recent) birth is an important factor indirectly affecting earnings. The older at f irst birth, the younger the youngest child at the survey date, the fewer hours the mother will be working, and, as a result, the less she will earn. Thus Boo and B ilsborrow found that later childbearers, among whites, actually earned less, but this was because they worked less. Spouse' s or Other Family Income Again, results are consistent. Age at first birth has no direct impact on other family income (Xoo and Bilsborrow, 1979; Bofferth and Moore, 1979; Card, 1977; Haggstrom et al., 1981) . Among males, arc five years out of school adolescent fathers were earning more than compar- able peers; 11 years out the difference had disappeared (Card, 1977J. At that point, they were all out of school and in the labor force. There are a number of indirect ef feats. An early f irst birth is assoc fated with less schooling completed at age 27, which is associated with lower income of other family members at age 27 (Hof forth and Moore, 1979~. An early first birth is associated with having a large number of children, which is associated with a lower income of other family members at age 27. Because of these two effects, an early f irst birth is associated with lower income of other f Emily members at age 27, but the effect is indirect. Family Income, L iving Standards and Poverty The effects of age at first birth on income and poverty are consis- tent with its effects on a female respondent's own earnings and other family income. There is no direct effect of age at first birth on family income, net of other factors (Koo and Bilsborrow, 1979; Hofferth
136 and Moore, 1979; Haggstrom et al., 1981) . Nor is there a direct impact of age at f irst birth on whether or not the family is poor (Koo and B ilsborrow, 1979; Hofferth and Moore, 1979) . In contrast, early childbearers have higher living standards in midlife {age 3S-44} because they have fewer Equivalent adult consump~ tion units. (EACs) (i.e., fewer children in the bc~ne) than later child- bearers (Koo and B ilsborrow, 1919) . Although early childbearers had a greater number of children than later childbearers, they had them a longer time ago. Thus by the time the mother reaches age 35-44, most of the children of early childbearers have grown up and left home In contrast, the children of later childbearers are younger and the majority still remain in the home. This Mints out the necessity of comparing young women who are at similar points in the life cycle to be able to make adequate comparisons of economic well being. Comparisons at a later point in the life cycle would be useful. (For a comparison of delayers with average age childbearers at a much later point in the life cycle see also Hofferth, 19841. Indirect Effects of Early Childbearing Even though there is no direct effect of an early f irst birth on family income or poverty status of young women, it is clear that there may be substantial indirect causal effects due to the impact of an early bi rth on schooling and on family size and composition. Level of schooling is a consistently important factor determining earnings. Family size is a consistently important factor affecting labor force participation by the mother and per person availability of income. Therefore both variables can be expected to affect family inccae and poverty status of a mother by af fecting whether or not she is employed and how much she earns. And since both are affected by an early first birth, an early f irst birth will indirectly affect later family income and poverty. By tracing out these intervening paths we can better identify the kinds of impacts that an early first birth has, the magni- tude of each of the effects, and the overall contribution of an early f irst birth to economic well-being . There are two studies (Koo and 8ilsborrow, 1919 and Hofferth and Moore, 1979) that have traced out a complex chain of effects from a first birth to later family income and poverty. These two papers form the basis of this part of the review. Other papers that have looked at part of the process will be referenced when appropriate. Two analy- ses were conducted in each study: one on all women ; a second on only those women who had a f test bi rth before they reached age 19.
137 Results for all When According to data for women of all races from the National Longitu- dinal Surrey of Young Women. for each year a first birth is delayed, other family income at age 27 increases by almost S500 per year; the womants own income increases by 8200 (Hofferth and Moore, 1979~. The effect is stronger for whites than for blacks. The effects are similar for whites in the NSFG . The effect of delaying a bi rth f roen 17 (or under) to age 18-19 is to raise f Emily income by almost S700 (Koo and B ilsborrow, 19 79) . As a result, for each year a young woman delays her first birth, her chances of being in a family below the poverty level is reduced by 2.2 percen~cage points among women of all races, a reduction over the total probability of being poor of 22 percent (Hofferth and Moore, 1979) . Both studies (Roo and Bilsborrow and Hofferth and Moore} found that, among women of all ages, the largest part of the indirect effect of an early birth con later economic well-being is due deco the larger family sizes of early childbearers. Among women of all races, over half of the impact on own earnings and 80 percent of the impact on poverty status is due to differential family size in the Hofferth-Moore study. Twenty percent of the total impact on own earnings is due to the impact of an early birth on work experience and on hours worked last year. Only 6 percent of the total effect of an early first birth is through schooling. Of the effects on other family income, three quarters is due to the smaller families of delayers, one-quarter to greater schooling. Finally, of the total effect on poverty, 80 per- cent is due to smaller families of postponers, 12 percent to ~ rester schooling, with 8 percent to differential labor force participation. In the Koo~Bilsborrow study, among women of all ages, the largest por~cion of the indirect effect of an early birth is also due to the dif ferential family sizes of early and later childbearers. One of the reasons is that a path through education was not specified for the total sample of women. But even when a path through schooling is specif led, the effect through family size is as large as that through schooling . It is certainly clear, therefore, that among women of all ages, the effect of a first birth through education on later earnings is very small, while that through family size is substantial. Adolescent Childbearers It is among the very earliest childbearers that we would expect the largest indirect effects of childbearing and the largest impact through schooling. The total effect of delaying a f irst birth for one year during the teen years on the earnings of the youngest childbearers is larger than that among women of all ages (Hofferth and Moore, 1979).
138 Seventy percent of the impact of early birth on own earnings of those whose f irst child is born at or before age 18 operates through reduced schooling. Another 30 percent operates through number of children. Roo and B ilsbor row also f ind a strong ef feet through schooling for white teenage childbearers, but not for black teen childbearers. About half of the total effect is due to reduced schooling. Blacks versus Whites In both studies the results are weaker for black than for white women. Age at first birth does not appear to be as important for the black woman as it is for the white woman. Among both black and white women the pr imary negative indirect impact of an early f irst birth on later economic well-being is through its impact on family size. An early first birth means more children by age 21 with its concomitant negative impact on labor force participation and earnings (Bofferth and Moore, 1979~. However, among black women, early childbearers accu- mulate more work experience than later childbearers, increasing their earnings at age 27. Thus an early f irst birth is associated with some what higher well-being among blacks; among whites, early childbear ing predicts substantially lower income. An early first birth has no imp pact d irectly or ind irectly on the incomes of other funnily members and very little on the probability of being poor among blacks, whereas there is a substantial negative impact of an early f irst birth among whites both on other family incomes and on the probability of being poor at age 2 7. Welfare Receipt Early childbearers are more likely to be in households receiving AFDC, but the relationship is mostly indirect. Once other factors such as socioeconomic background, education, age at first marriage and tom ing of first birth are controlled, the relationship disappears (Hove et al., 1978~. A premarital first birth is associated with welfare receipt, particularly among young female heads. A premarital birth inc reases the probability of going on welfare for those not enrolled and reduces the probabil ity that those already enrolled will exit wel- fare {Moore at al., 19781. Another way to look at the problem is to ask whether early child- bearers are disproportionately represented among welfare recipients. It appears that they are. Moore { 1918~ aproached this question by asking what proportion of AFCC and non-AFDC households contain mothers who began childbear ing as teenagers? She ~ and ocher researchers; found that in the mid 1970s between 60 and 80 percent of mothers under 30 in AFOC households were teen mothers, compared to only 35 percent of mothers in non-AE DC households {Moore, 1978; Moore and Burt, 1982; Block and Dubin, 1981; Scheirer, 1983~.
139 There appears to be an a. - ciation between early childbearing and welfare receipt. However, this effect is mostly indirect: an early pregnancy may precipitate ~ premature and instable marriage. An early and premarital birth creates a family form with a high probability of needing public assistance. The low educational levels and large family sizes of teen childbearers increase their probability of depending on public assistance later on. Factors Leading to Successful Early Childbearers A recent study (Furstenberg and Brooks~Gunn, 1985) has explored the factors associated with variation in outcomes among early childbe~rers. The researchers followed up a sample of 300 women in Beltmore who had their f irst child at age 18 or younger one, three, f ive and seventeen years after that first birth. The purpose of the study was to see what factors and conditions affected the adep~cation of the early child- bearers and their eventual economic well-being. The outcomes, measured 17 years after f irst birth, were l) whether receiving welfare in 1984, and 2) whether economically secure, that is whether family income totalled S25, 000 per year or more in 1984. The factors associated with whether a family was econcnically secure in 1984 were almost identical to those associated with whether a family was receiving welfare, though the direction of effects was the reverse. Three family resource factors were associated with later economic well-being: high parental education, small parental family size and welfare experience as a child. Those whose parents had high levels of schooling were twice as likely to be secure a. adults, and 4 times less likely to be on welfare. Those who came frae smaller families were more likely deco be secure and less likely to be on welfare because they were less likely to have a second child soon after the first. Finally, those from welfare families were more likely to receive welfare them selves soon after a birth, and as a result, were more likely to receive it and less likely to be economically secure as adults. These are face tars over which the individual has relatively little control. Characteristics of The individual during the high school years and over which some control can be exercised include school performance, school continuation, type of school attended and educational aspira- tions. Those who had high aspirations were more likely to to attend the special school for pregnant girls and to remain in School, both of which were associated with ~ lower likelihood of being on welfare and a greater chance of being economically secure later on. Being at grade level was also associated with a greater chance of being economically secure as an adult. The factor over which individuals have substantial control is their use of birth control. The researchers found that those who used birth control had fewer add itional children soon after the f irst, and were more secu re and less l ikely to be on welfare later on as a result.
140 Planned interventions were also important. Girls who attended either the special school for pregnant girls or attended the hospital prenatal program were more likely to contracept than those who didn't (Furstenberg and Brooks - ;unn, 19851. What career contingencies, factors impinging during the young adult years, affect later well-being among teen mothers? Those who married and who remained married were less likely to receive welfare and Snore likely to be economically secure 17 years after the first birth. Mar- r i age i s a key to economic success, but only when the mar r iage lasts. Unfortunately, the chances of having a stable marriage were very low in this sample. Early marriage was usually a losing bet. Home n who married early were especially prone to economic dependency when their marriages did not work out because they f requently had chic short their educ^- tional careers to enter matrimony. Women who married later, espy cially if they did not wed the child's father, were also in a pre- carious situation for these relationships were particularly prone to dissolution. Women who delayed marriage indefinitely to con- tinue their education usually avoided economic dependency but they rarely could achieve economic security on the strength of their own earn ing powe r (Fu r stenbe rg and B rooks~Gunn, 198 S: 92 ~ . Work exper fence appears to have little impact on economic success. In fact, early work experience may be harmful, particularly if it prep vents school completion (Furstenberg and Brooks~Gunn, 19857. Residen- tial exper fence has a small impact on economic success. Women who remained in the parental household for three or more years were less. 1 ikely to be economically secure at ~ he 17 year follow~up, althoug h the effect is small. Thus, although sore parental support and help is im portent after a f irst birth, lengthy coresidence does not enhance echo nom ic independence . The most important factors in later economic success or failure were family resources, aspirations, marital success and control of fertility. Clearly early childbearers who are ambitious, who continue in school, who use birth control and who avoid a rapid subsequent birth are better able to control their long run family size. The earlier rem suits show that this is one of the most important ways that early childbearers can increase their prospects for economic securi~cy and independence as adults. SOC IETAL COSTS Early childbearing has an impact on society, for when individuals cannot realize their full educatione1 and occupational potential, son ciety loses their economic contributions. In addition, if early child- bearers utilize public services more than other women, public expendi- tures on programs such as AID to Families with Dependent Children ~ APDC I, Med icaid , and food stamps inc tease.
141 The precarious section has shown that AF0C mothers are more likely to have been teen mothers than are American women in general. Tree studies have attempted to estimate the cost of teen childbearing in terms of the public expenditures on women who were teens when they had a f irst birthe This total does not necessarily represent the amount that could be saved if all these mothers had postponed their first birth, since some would have required public assistance regardless of their age at first birth. Moore (19t8) and Wertheimer and Moore (1982) analyzed three data sets to estimate 1975 welfare expenditures on teen mothers. The rem suits show that about half of the AEDC budget goes to households in which the mother was a teenager at f irst birth, about 84. 65 to 85 billion in expenditures just through AFDC (Moore and Burt, 1982:Table 8) . Adding food stamp benef its plus Medicaid benef its to mothers and children increases the total to S8. S5 billion in 1975 {Moore and Burt, 1982 :Table 9) . Scheirer ( 1982~ estimated AFDC payments to current and prior teen mothers under age 30 (using the 1975 and 1971 AFDC surveys} to total S2 . S billion in 1975 and S3 billion in 1977. Moore' s estimate of pay- ments to households of women age 14-30 and who gave bi rth before age 20 was S2.4 billion in 1975. The estimates based on a number of dif- ferent data sets are very similar. Block and Dubin estimated APDC costs for teen childbearers in Monroe County, New York in 1977 and 1978e They found the average cost per case to be S4,262 and S3.494 in 1978 for teen and non-teen child- bearers respectively under 30 in that year. Scheirer also found that households of teen mothers received larger grants: however, this was because of the larger number of children of teen childbearers than older mothers. Once other factors were controlled the direct effect d isappeared. Block and Dubin showed that over time older childbearers do catch up somewhat; however, substantial differences in family size remaine Scheirer also showed that the length of time on welfare is a function of age at f irst birth. Early childbearers spend slightly more time on AFDC. Thus the higher welfare cost of early childbearers is due to three factors: the higher proportion of early childbearers who are recipients, the higher cost per case, and the longer duration of payment (Scheirer, 1982:31. Finally, in a recent study {see this volume, Chapter 10, using a similar mehtodology to that of Moore (1978) Burt estimated total AFOC costs in 1985, due to teenage mothers, to total 16. 6 billion dollars, doub le the 197 5 Moore estimate . THE HYPOTHETICAL IMPACT OF POLICY INTERV=TIONS Fu ether analyses addressed the relative impact on public sector costs of reducing births as opposed to mediating the effects of an
142 early birth {Moore and Wertheimer, 1984 :Tables 1,2; Wertheimer and Moore, 1982:Table 37~. Three scenarios which reduced first births to teens and three which would mediate the effects of an early birth were compared to a baseline scenario in which present trends were continued. The results show dollar savings for all approaches, but a much greater savings when a f irst b irth is averted. The g Neatest savings occur when the fertility of all teenagers is reduced by 50 percent--the number of women age 20-29 receiving AFDC payments in 1990 would be reduced by 35 percent, compared with the baseline scenario; public sector costs for AFDC, Med icaid and Food Stamps for families of women 20-29 would be rem duced by an est imated S 1.4 bill ion. Eliminating births to unmarried women under 18 reduces the number of Amen 20-29 receiving AFDC by 17 percent and reduces public sector costs for them by S. 9 billion. Reducing the fertility of teens under age 18 by 50 percent reduces the number receiving AFDC by 14 percent and reduces public sector costs by S 72 billion. Reduc ing the subse- quent childbear ing of young teen childbearers reduces by 11 percent the number receiving AFDC, and reduces public sector costs by S1 billion. The reduction in the number receiving AFOC due to reducing school drops out of teen childbearers and to increasing their marriage probabilities are two and 11 pe scent, respectively . These represent declines in ex- penditures of S.22 billion and S. 77 billion. Thus the results support the common sense notion that prevention is preferable to remedial cures. Of the ameliorative strategies, reducing subsequent fertility is the most effective, and the one that appears to become even more signif icant over time. The scenario with the least impact is reducing school d Copout. Although initially surprising, this result seems to a r ise f ram the relatively low economic return to education for women such that even well-educated women earn relatively little. Marriage appears to imp rove the short-term economic status of young women more then add itsona1 school ing . Con~crary to initial expectations, none of the scenar ios has a sig- nif leant impact upon labor force participation, hours worked, earnings or taxes. There are several possible reasons for the lack of effect. Moore and Werthe~mer (1984; Wertheimer and Moore, 1982) cite as reasons the lack of strong relationship between education and occupational attainment for this group of young women. This argument is supported by data from McLaughlin, 1977, who finds that early childbearers are less able than later childbearers to translate additional schooling into ~ rester work exper fence and higher earnings potential. Fursten- berg and Brooks~unn (1985) also failed to find a strong relationship among early childbearers between schooling and later economic secur ity as adults. However, an alternative explanation is the differential life cycle stage hypothesis referred to earlier. The women in the we rthe imer-Moore study are st ill relatively youns--ages 20 to 29 in 1990, the endpoint of the computer simulation. S. ince this is the period of childbearing for most women, delayers would be beginning families at the tine that the early childbearers would be moving back into the work force. This would tend to minimize d if ferences between early and later childbearers.
143 It is important to note that differential patterns of childbearing do have a ve ry st rang impac t an publ ic sector costs even at ages 20 to 29. The research cited above strongly suppports the previous conclu- sion that early childbearing does have substantial long term economic costs for both the individual and for society, and that rapid subse- quent childbearing and large family sizes among early childbearers are a ma jor reason for the g rester disadvantages of early childbearers ana the la rge cost to the publ ic. These ser ious consequences underscore the benef its of policies which delay the first birth and prevent or delay subsequent births to teenagers. Scheirer et al. ( 1982) found that the indirect effects of a one- year increase in mother's age at first birth, aggregated across the total population in 1975 of AFDC families with a mother under age 30, generated expected costs savings of S12. S million per month or approximately S1S0 million per year "without including any savings generated by any reduction in the number of rec ipients. SUMMARY AND COliCLUSI ONS This review has included only those studies that controlled for several important pr for d if ferences between early and later child- bearers, of which socioeconomic status background is the most impor- tant. Several studies were able to control for aptitude as well--the Card study, for example. All the studies cited are consistent in at least one regard. All f ind an additional negative impact of early childbear ing on later economic well-being af ter ad justing for back- g round and other prior d if ferences. The studies reviewed here are especially important because they reveal the process whereby an early birth affects later economic well- being. First, most of the impacts on later economic well-being are indirect. That is to say, an early birth reduces schooling and in- creases later family size. ~ t is these variables that reduce later labor force participation, earnings and family income, not the early bi rth per se. This implies that if the links between an early bi rth and schooling or family size could be broken, so would the link be tween an ear ly b i rth and economic d i sadvantage . Th is is the opt i- mistic part. It has proved difficult, in fact, to break these links. More research on the factors associated with lessening these con- nect ions is needed . Second, the factors that disadvantage early childbearers relative to later ch ildbearers in economic well-being are the same factors that discriminate the more f rom the less successful early childbearers. One difference is that for certain types of adolescent programs eligibility depends on childbearing status. These studies have also pointed out important race and ethnic dif- ferences. Because so little is known about Hispanics, this chapter
144 focuses on black-white differences. The important difference is that blacks are not affected as negatively by an early rather than later f irst bi rth as are whites. There are several possible hypotheses as to why this is so. F irst, early childbearing it Common in the black community; therefore, institutions and mechanisms have developed to help young women cope. A second hypothesis is that opportunities have not developed enough in the black community so that the differences among young women with high and low opportunities are not as great. Another hypothesis might be that blacks start earlier, but that they terminate childbearing earlier; thus they can devote themselves to employment in their early twenties, when later childbearers are just beg inning . Another hypothesis is that the reservat ion wage for blacks is so much lower than whites that they do not have the luxury of rem mailing out of the work force as do whites. Although most of the research conducted to date has analyzed the impact of an early f irst birth on the young mother, the evidence pre- sented suggests important impacts on the father as well. More research needs to be conducted to better describe the impact of early father- hood on young men. Improved data are just now becoming available ( see, for example, Marsiglio, 1986) and should increase researchers' ability to determine the consequences of early childbearing for males. A f inal point is that all the studies mentioned here are based on data collected in the 1960's and early 1970' s. Birth years of the respondents date f ram the late 1920 ' s to the early 1950 ' s. Thei.r high school experience predates the implementation of Title IX in 1975, pro- hibiting discrimination against pregnant or parenting teenagers in publicly funded school programs. Thus we don't know what changes have occur red between these stud ies and cur rent students. There are now enough years of longitudinal data available from several recent national data resources to replicate some of these studies of long te rep consequences of teenage childbear ing and see what changes have occur red . Of course, as adolescents charge, the rest of sac iety has also been changing. On the Oh-~- hand, today family size remains low and education high. Relative to the ma jar ity of adults, not completing high school and having more than two children probably represents an even ~ rester d isadvantage than it might have been even one decade ago . On the other hand, seve ral stud ies show that a small amount of add i- tional schooling would decrease early childbearers' dependence on public assistance and increase their economic security as adults only slightly relative to the large impact of a change in childbear ing patterns. Thus one conclus ion is that although increasing school completion is an important ob jective, the relationship between schooling and women' s earning power is still too weak for the latter alone to raise living standards. Even today women' s long tern, economic secur ity is heavily dependent on mar ital success and fertility control.
