Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Assisting Young, Unmarried Mothers to Become Self-Sufficient: The Effects of Different Types of Early Economic Support Author(s): Jodi R. Sandfort and Martha S. Hill Source: Journal of Marriage and the Family, Vol. 58, No. 2, (May, 1996), pp. 311-326 Published by: National Council on Family Relations Stable URL: http://www.jstor.org/stable/353498 Accessed: 13/06/2008 16:33 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=ncfr. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected]. http://www.jstor.org JODIR. SANDFORTAND MARTHAS. HILL University of Michigan Assisting Young, UnmarriedMothers to Become Self-Sufficient: The Effects of Different Types of Early Economic Support This article examines a sample of young, unmarried mothers from the Panel Study of Income Dynamics (PSID) and considers how different types of economic support received soon after their first child is born contributes to the later self-sufficiency of young, unmarried mothers. It expands conventional categories of income support-AFDC, food stamps, child support-to include shared housing and relatives' assistance. The model also contains various behaviors of young mothers after the birth of their first child. The findings suggest that certain economic supports assist these mothers and that life choices they make after their child's birth are important to self-sufficiency. Recent Congressional proposals to reform the welfare system have focused attentionon strategies for discouragingout-of-wedlockbirths,particularlyamong young parents.This attentionis fueled by the belief that pregnancy outside of Institutefor Social Research,University of Michigan, Ann Arbor, MI 48106-1248 [email protected] hillm@isr. umich.edu). Key Words: income support, self-sufficiency, teenage mothers, welfare. marriageis the cause of manyothersocial ills. As Charles Murray(1993), a neo-conservativeresearcherwrites: "Illegitimacyis the single most importantsocial problemof our time-more importantthancrime,drugs,poverty,illiteracy,welfare,or homelessnessbecauseit driveseverything else." Some policymakersseem to have adopted this view, positingthatif they reformthe welfare system there will be fewer unmarriedwomen having children,which, in turn, will reduce the othersocial ills. Yet social science evidence contradictsthese beliefs. Researchdocumentsthat income support from welfare programssuch as Aid to Families with DependentChildren(AFDC) is not a cause of women bearing children outside of marriage (Duncan & Hoffman, 1988; Ellwood & Bane, 1985; Moffit, 1992). Instead, a host of demographic and social trends are producing lower rates of marriage,higher ages at first marriage, and greaterincidence of divorce and separation. Marriedwomen of all races and classes are having fewer childrenat the same time that unmarried women are having more (Ellwood & Crane, 1990; Furstenberg& Brooks-Gunn,1986; Mare & Winship, 1991; Wojtkiewicz,McLanahan,& Garfinkel,1990). While reformingthe welfare system is likely to do little to changethese complexdemographic Journalof Marriageandthe Family58 (May 1996):311-326 311 Journal of Marriage and the Family 312 trends, policymakers have seized on reform as a panacea and are paying scant attention to ways that reform could address other desirable goals, such as assisting low-income families become self-sufficient. Although policymakers rhetorically have expressed their concern for increasing self-sufficiency among the poor, little attention has been given to defining self-sufficiency or considering how it is best achieved. This research seeks to address both public concern about outof-wedlock childbearing and the desire to increase the self-sufficiency of young mothers. Using a nationally representative sample of young women who were unmarried when they gave birth to their first child, we examine how economic assistance helps women in this vulnerable position achieve economic self-sufficiency later in life. We examine existing types of economic support-both public and private-and predict how these sources of support affect young, unmarried mothers' chances of becoming economically self-sufficient by the time their first child enters school. We also seek to understandthe process of achieving self-sufficiency by examining a number of factors that could mediate the effect of assistance on later self-sufficiency. PRIORRESEARCH Consequences of Adolescent Childbearing Much research has been done in the last 20 years on the consequences of adolescent sexual behavior, pregnancy, and parenting. (See Hayes, 1987; Miller & Moore, 1990; and U.S. Department of Health and Human Services, 1995, for reviews.) Although adolescent childbearing was at its peak during the baby boom, most young mothers at that time were married. Public attention to adolescent pregnancy grew, however, when the proportion of births to unmarried young women grew. By 1993, 72% of the births to teenagers were to unmarried women (U.S. Department of Health and Human Services, 1995). Public concern about early parenting is understandable because there are a number of negative consequences for the young mothers and their children. Teenage women who have children are less likely to graduate from high school or get a high school equivalency degree than women who delay childbearing until their 20s (Geronimus & Korenman, 1992; Hoffman, Foster, & Furstenberg, 1993; Mott & Marsiglio, 1985; Upchurch & McCarthy, 1989). Teenage mothers are more like- ly to be poor, to have low occupational status, and to be dependent on public assistance. In addition, they are less likely than other adolescents to marry later in life and, if they do marry, to have stable marriages (Furstenberg, Brooks-Gunn, & Morgan, 1987; Geronimus & Korenman, 1992; Hoffman et al., 1993; Kellam, Adams, Brown, & Ensminger, 1982; Maynard, 1995; Teti & Lamb, 1989; U.S. Department of Health and Human Services, 1995). Although the size of these effects in conventional estimation approaches may be upwardly biased, even recent research using fixed effects models of sister pairs to more fully control for family background finds statistically significant effects of teen birth (Geronimus & Korenman, 1992, 1993; Hoffman et al., 1993). The Concept of Self-Sufficiency Because young, unmarriedmothers are at risk for a host of negative life events, they can use assistance. A desirable goal is to provide that assistance in a way that fosters, rather than discourages, selfsufficiency in the long run. Although formal education was thought to be the path to increasing self-sufficiency, recent evaluations of welfare-towork programs suggest a more complex and variable process, with work experience playing an important role in developing the skills needed to leave welfare (Martinson, 1995). Although self-sufficiency is frequently heralded as the goal of public assistance programs, there seems to be no research that systematically explores its definition or the process of its attainment. This ambiguity in defining the goal prevents us from achieving it. For help in understanding the concept of self-sufficiency, we turn to sociological and economic literature that examines intergenerational transmission. This literature considers how parents' characteristics influence the achievements of their children in adulthood, including educational attainment, occupation, and income (Corcoran, Gordon, Laren, & Solon, 1992; Haveman, Wolfe, & Spaulding, 1991; Hill & Duncan, 1987; Sewell, Hauser, & Wolf, 1980; Solon, 1992). This literature helps inform the models used in this research by identifying how characteristics of family background directly and indirectly influence these outcomes, by pointing to factors that mediate this effect, and by suggesting that the predictors of self-sufficiency are dynamic and can be found throughout the life course. The intergenerational transmission literature, however, does not consider self-sufficiency Self-Sufficiencyfor Young Mothers as an outcome and focuses mostly on male samples. This article is a step toward better understanding the process of how young, unmarried mothers achieve self-sufficiency. Sources of Economic Support We are concerned with understanding how the sources of economic support received earlier in life help predict self-sufficient income later in life. As such, this project grows out of a traditionof social science research concerned with how the source of income influences its effect. The clearest example of this in the literature is the distinction between parental welfare income and parents' earned income in empirical models of young adults' attainments (Corcoran et al., 1992; Corcoran & Adams, 1993; Duncan & Yeung, 1995; Ellwood & Bane, 1985; McLanahan, 1985). Scholars make this distinction between these two types of income to test the hypothesis that income from welfare influences the achievements of children differently than income earned in the labor market. Although the results of these studies are somewhat mixed, they suggest that welfare income exerts different effects than earned income on at least some groups of children. Research on the association between childsupport payments and children's educational achievement (Knox & Bane, 1992) also backs up the hypothesis that the source of income is important. This research shows that larger amounts of supportpaid by a noncustodial parent are associated with children completing more years of schooling and with both a higher likelihood of graduating from high school and a higher likelihood of entering college. Other research (Hill & Duncan, 1987) shows that an assortment of types of family income-father's earnings, mother's earnings, assets, and welfare programs-serves differentially as predictors of children's completed education and later wage rates. This leads us to hypothesize that different types of income support have different influences on the later self-sufficiency of young mothers. Because our proposed study deals specifically with young, unmarried women, we broaden our concepts of economic support from the more commonly used welfare income and child support to include financial assistance from relatives and shared housing support. Past research on young, single mothers indicates that families help to cushion the effects of early parenting in many ways (Caldwell, Antonucci, Jackson, Osofsky, & 313 Wolford, 1995; Furstenberg & Crawford, 1978; Hogan, Hao, & Parish, 1990; Presser, 1980). Direct financial support from family members is one way of assisting young mothers. A nationally representative study found that 25% of African American single mothers and 38% of White single mothers received at least half of their income from other family members (Hogan et al., 1990). The importance of families in boosting the income of young mothers is reinforced by additional studies as well (Caldwell et al., 1995; Furstenberg & Crawford, 1978; Presser 1980). The creation and maintenance of intergenerational households is another important way that parents help support their young, pregnant daughters (Furstenberg et al.,1987; Goldscheider & DaVanzo, 1985; Hill, 1990; Scheirer, 1983; Trent & Harlan, 1990, 1994). Sharing housing with family reduces the living expenses of young mothers and may allow them to work outside the home or continue their formal education. It also may decrease the likelihood that young mothers will quickly have additional children (Presser, 1980). As we will see, altering these types of behaviors may directly increase their long-term self-sufficiency. There are other types of support young mothers can receive that may affect their later ability to become self-sufficient. Chief among these are no-cost child care, provided either by grandparents or friends (Hogan et al., 1990; Presser, 1980) and social and emotional support that can influence motivation and depression (Caldwell et al., 1995; Presser, 1980). These factors, however, cannot be included in this project because the data set used in this analysis, the Panel Study of Income Dynamics, does not contain such measures. Factors Influencing How Economic Support Affects Self-Sufficiency There are clearly ways that women can react to these various types of economic support that will improve their later self-sufficiency. The support may facilitate participation in the labor market or continued education, both of which enhance wage rates. The support also may influence chances of meeting a potential spouse with a high wage level, as well as chances of giving birth to additional children. Work experience can help young women build the qualifications they need to secure better paying jobs in the future if they are able to make strong labor market commitments. Previous research shows that it is difficult for young mothers to get 314 work experiencein the first years following the birth of a child. In a national study of young mothers,both marriedand unmarried,Ortiz and Fennelly (1988) found that 2 years after giving birth,only 40%of Whitewomen,30%of African Americanwomen, and between20% and 40% of Hispanic women were working outside their homes.No grouphadreturnedto its pre-pregnancy level of participationin the labor force. These percentagesare probablyeven lower among unmarriedmothers.Evenif they areableto get work experience during the first years after a child's birth, young mothersoften find that their childcare responsibilitiesrestricttheirabilityto obtain the typesof jobs thatwouldinvolvethe trainingor experiencethatwould subsequentlylead to higher wages.The GeneralAccountingOffice (1991) has foundthateven whenyoung,single motherswork full-time,most remainnearor below the poverty level. The GAO concluded,"If mother-onlyfamilies are to escape from poverty, the majorityof poor, single motherswill need eithersupplementaryincomesourcesorjob trainingthatraisestheir earningssubstantially" (p. 8). Considerableresearchhas documentedthe importanceof educationto latereconomic viability (Ashenfelter& Krueger,1992; Danziger, 1991; Corcoranet al., 1992). This suggeststhatsupport that enables continuededucationwill contribute to the laterself-sufficiencyof unmarriedmothers. The income that young mothers receive from AFDC and the supportthatthey get from sharing housingwith theirparentsmay help them stay in or returnto school. This, in turn, may enhance wages and thus allow them to earn higher selfsufficientincome by the time theirfirst child entersschool. Earlyeconomic supportmay also help young mothersbuild connectionswith potentiallyeligible marriagepartners,and marriage,in turn,provides access to an importantincome source-a husband'sincome.Marriagemay be only a shortterm solution to women's poverty, however, because very real possibilities of separationor divorcemake even marriedwomen vulnerableto economic hardship.Existing research has produced mixed results regardingthe influence of marriageon eventual self-sufficiency of young mothers.Workby Teti andLamb(1989) suggests thatmarriagemay not be a viablepathwayto selfsufficiencyfor young mothers.Yet Brooks-Gunn and Furstenberg(1987) arguethat a stable marriageis criticalto long-termeconomicviability. Journal of Marriage and the Family A final factorthatcould influencethe relationship between economic support and selfsufficiencyis whetheror not a young womanhas additionalchildrensoon afterthe birthof her first child (Furstenberg et al., 1987). Research by Presser (1980) suggests that women who share housingwith their parentsare more likely to use birth control and to terminatetheir pregnancies becausethey don't wantto jeopardizethe support they receive.Parentsmay be less likely to continue sharing their homes if unmarrieddaughters continueto have morechildren.In addition,there areclearlyeconomiccosts associatedwith raising more children,particularlyif the motheris doing it alone. RESEARCHPROCEDURES Data and Sample The data for this analysis come from the Panel Study of Income Dynamics (PSID), a nationally representative,longitudinaldata set that has followed individualsand families since 1968 (Hill, 1992). The PSID initiallyoversampledpoorfamilies, but constructed probability of selection weights that adjustfor the oversampling,as well as for nonresponse.We have drawnour sample from the full PSID sample and then used these weights in both our univariateand multivariate OLS regression estimates to allow inference to the U.S. population.Any missing values in this dataset were assignedthe meanvalue of the variable or a value appropriate to the population. Oursampleconsistsof 302 youngwomenwho had theirfirst child when they were betweenthe ages of 16 and22, andwho were unmarriedat the time of the birth. In order to have measuresof their backgroundcharacteristics,membersof the samplewere observedin the PSID duringadolescence between the ages of 12 and 14. They also were observedwhen theirfirst child was between the ages of 5 and 7-the time of observationfor this analysis. Altogether, each member of the samplehas between11 and 19 yearsof datain the PSID drawnfromthe 1968-1990 interviewyears. Each had her first child sometimebetween 1972 and 1983. Measures and Model In this analysis,we are examiningthe amountof self-sufficientincomethe youngmotherhas when her first child is enteringschool. Becauseof limi- Self-Sufficiencyfor YoungMothers tationsin the availabilityof data,we arerestricted to examiningthis outcomeno laterthanwhen the firstchild is age 7. As time passes,moredatawill become availablein the PSID, and this outcome thencan be assessedwhenthe childrenareolder. For the purposeof this analysis,we have operationalized self-sufficient income in two ways. First, we examine how our model predicts women's self-sufficient income, defined as the young mother's average annual income from labor,child support,andassistancefromrelatives. This measure is a strict representationof these women's self-sufficient income because it excludes income from AFDC, food stamps, and fromanymenthey subsequentlymarried.The outcome measuredoes, however,include assistance from otherrelativesbecause this type of income transferis connectedto the woman,herself,andis notdirectlydependenton hermaritalstatus. Secondly,we examinehow ourmodelpredicts anotherindicatorof self-sufficientincome-what we call marriage-oriented self-sufficientincome. This outcomeis adjustedto reflect the changing economic conditionsof women who marryafter the birthof theirfirst child. It is definedas average annualincome in the woman's family from her labor,her husband'slabor,child support,and assistancefrom relatives.While this measureaccounts for all privateresources available in the of family, it may be a less reliablerepresentation these mothers'economicprospectsbecauseof the high prevalenceof unstablemarriagesamongthis population.For both outcomes,self-sufficientincome was averaged over 3 years when the woman's first child was aged 5-7 because such measuresaremorestablethansingle-yearestimations of income. In this analysis, we are principallyinterested in how the differenttypes of economicsupportreceived by young motherswhen their first child was betweenages 1 and 4 influencetheir subsequent self-sufficiency. The sources of the economic supportare both public sources (AFDC and food stamps)and privatesources(child support, sharedhousing, and financialsupportfrom relatives). In our analysis, these and all other measuresof income are expressed in 1990 dollars. In additionto these predictors,the model includes a numberof controlvariables.First,there are two variables-county unemploymentrate andcountyunskilledwage rate-that measureenvironmentalconditionsat the time of the outcome that could influenceeconomic well-being.Then, 315 there are a numberof measuresof the mother's family background,all observedduringa 3-year period when she was between 12 and 14 years old: herrace,whethershe lived in a ruralarea,the number of years she was poor, the average amountof AFDC income her parentalfamily received, the number of other children in her parentalhousehold,the numberof yearsshe lived with one parent,her mother'slevel of education, and a measureof her expectededucationalattainment.The variablerepresentingherexpectededucationalattainmentcomes from data collected in the early years of the PSID when parentswere asked whetheror not their childrenwould complete high school. All of these characteristicsof parentalfamily backgroundare included in our model because previousresearchshows they directly or indirectlyinfluenceeconomic outcomes (Bumpass & McLanahan,1989; Furstenberget al., 1987; Geronimus& Korenman,1992; Hogan, 1986). Our third set of predictorsmeasurescircumstances that existed just before the birth of the child. These variablesare includedas controlsfor differencesamongthe young motherswhen they became pregnant.Measures of her parents' income-to-needs ratio (e.g., their poverty status), her age at time of birth, and the state's AFDC guaranteeduringthe year before her first child was born were included. Women not living in theirparents'homes had missing data on the income-to-needsvariable.A dummyvariablewas createdandincludedin all regressionequationsto control for this. The state AFDC guaranteewas includedto explorewhetheror not the level of the state's AFDC grantinfluencedthe mother'sselfsufficiency. We also included a measurein the tables called "statesharepenalty"that represents whetheror not the statehad a policy thatreduced the level of AFDC benefits for families who sharedhousing.This penaltycouldhave servedas a disincentive to accepting housing assistance from relatives and, hence, may have influenced the receiptof this formof support. Taken together,the outcome variables,focal predictors,environmentalconditions,background of the mother,andinitialconditions characteristics just priorto the birthof the firstchildcomposethe base model of this analysis.To this base model, we add variables representingeducation, work, and family behaviorsthat the young women engaged in after the birth of their first child. The amountof workexperience,education,the length of a subsequentmarriage,or the numberof addi- 316 Journal of Marriage and the Family tional children may influence the amount of the self-sufficient income of these young women when their children are older. These variables are measured while the first child is between the ages of 1 and 4. We hypothesize that these variables represent various pathways through which economic support influences later self-sufficiency. For a detailed description of these variables see the Appendix. RESULTS Descriptive Analysis The weighted means of all variables included in this analysis are presented in Table 1. In this table (as is true throughout this analysis) all income variables, including those treated as outcomes, are measured in hundreds of dollars ($100s). For example, although the mean of women's selfsufficient income reads 63.44, the value actually is $6,344 annually. Likewise, the average value of income is marriage-oriented self-sufficient $16,191. As Table 1 reveals, the average annual amount of AFDC received after the birth of the women's first child was $1,871. The mothers annually received an average of $703 in food stamps, $293 in child support, and $131 in assistance from relatives. The mothers averaged 1.6 years out of 4 in which they shared housing with family members. Forty-one percent of this sample was Black, and 22% lived in rural areas while they were adolescents. The mean age at which the women gave birth to their first child was 18.5 years. In general, this sample of young, unmarried mothers came from disadvantaged backgrounds. Ninety-three percent of them lived in poverty for at least 1 year during their early adolescence (ages 12-14). Eighteen percent grew up in families in which their parents did not expect them or their siblings to finish high school. Although means provide a useful descriptive representation for some variables, a number of variables are more clearly illustrated by examining other aspects of their distributions. Additional descriptive statistics also serve to emphasize that counting or not counting husband's income as a part of self-sufficient income makes a vast difference. The distribution of the primary outcomewomen's self-sufficient income-reveals that 46% of the sample had less than $2,400 annually in self-sufficient income. Thirty-two percent had TABLE 1. WEIGHTED MEANS AND STANDARD DEVIATIONS OF MODEL VARIABLES Variables Outcomevariables Women's self-sufficient income Marriage-orientedself-sufficient income Early economic supports Aid to Families with Dependent Children Food stamps Child support Sharedhousing Financialsupportfrom relatives Environmentalcontrols County unemploymentrate Countyunskilled wage rate Conditionsin early adolescence WhetherBlack Always rural Years poor AFDC income Numberof siblings Years with single parent Mother's education Low educationalexpectations Conditionsjust priorto first birth Individual'sparentalincome-toneeds ratio State AFDC guarantee State sharepenalty Age at birth Year child born Education,work, and family behaviors Labormarketexperience Education Years married Additionalchildren Mean SD 63.44 73.13 161.91 144.09 18.71 7.03 2.93 1.64 1.31 24.31 9.74 13.78 1.47 4.03 7.37 5.61 2.90 1.04 .41 .22 21.97 4.10 .94 10.42 .18 .49 .42 .96 46.82 2.07 1.31 2.50 .38 .92 6.49 .34 18.47 78.40 1.15 2.72 .47 1.59 2.65 1.14 11.27 1.47 .62 1.19 1.55 1.69 .68 2.53 Note: All income variables are measured in hundreds of dollars ($100s). See Appendix for definitions of variables. self-sufficient incomes above the poverty line. In contrast, when we look at the second outcomemarriage-oriented, self-sufficient income-we see that only 24% of the sample had less than $2,400 annually. Sixty-one percent had self-sufficient incomes above the poverty line when their husband's labor income was included as selfsufficient income. Together with the results in Table 1, additional descriptive statistics show that this sample of young mothers exhibited diverse behaviors during the years when their first child was aged 1-4. Although we are unable to gauge the additional education received during that particular period, we find variation in the level of completed education, with a mean of 11.3 years but a modal value of 12 Self-Sufficiencyfor Young Mothers years (a high school degree). The data allow a clearer view of the women's labor market, marriage, and fertility behaviors during the 4-year interval following the birth of the first child. During that time, the young women gained an average of 1.14 years of full-time equivalent work experience. However, only 29% of them gained exactly 1 full-time equivalent year of experience, whereas 39% did not participate in the labor market at all, and 4% worked full-time throughout the 4-year period. The sample was strongly divided with regard to marriage and fertility behaviors. Fortynine percent married, 51% did not, and 50% had an additional child during that time, whereas 50% did not. Predictive Analysis-Women's Self-Sufficient Income In our analysis of factors predicting selfsufficiency, we first examine how, in multivariate contexts, early economic supports predict women's self-sufficient income. We then add education, work, and family behaviors to this base model to test the hypothesis that they act as potential pathways for how economic support influences later self-sufficiency. We consider how this same model predicts a marriage-oriented definition of self-sufficiency. Table 2 documents the construction of the multivariate model to assess the association between forms of economic support and later selfsufficiency. For column 1, self-sufficient income was regressed on the different types of economic support. In this preliminary model, child support and shared housing are both statistically significant with sizable positive coefficients. For column 2, we reran the equation, this time adding controls for environmental conditions at the time the outcome was measured and for women's family conditions during adolescence. Again, child support and shared housing remain statistically significant with sizable positive coefficients. For column 3, the estimation equation was further expanded by additional controls for conditions just prior to the birth of the child. Child support and shared housing again are statistically significant predictors with sizable positive coefficients. The adjusted R2 shows that this final model accounts for 26% of the variance in the outcome, a sizable proportion for microlevel analysis. Assuming this model is correctly specified, the results suggest that for every year a young mother shares housing with another family when her child is young, her 317 later self-sufficient income will increase by $1,000. For every $100 in child support she receives before her child is 5, her later selfsufficient income will increase by $168. The construction of this model also reveals that AFDC, food stamps, and financial support received from relatives are not statistically significant predictors of women's later self-sufficient income. Thus, while these sources of support do not bolster later self-sufficiency, neither do they hinder it. Adding the control variables does not appreciably alter the relationship between our focal predictors-the early economic supports-and later selfsufficiency. There are, however, direct relationships between some of the control variables and women's later self-sufficiency. First, the county unemployment rate has a statistically significant coefficient that is sizable and negative. This shows that women's self-sufficient income is inversely associated with the unemployment rate, lending support to Wilson's (1987) hypothesis that structural factors play importantroles in understanding why some people cannot find stable employment. The number of years the woman was poor in early adolescence also is negatively related to her later self-sufficiency, as is her parent's income-toneeds ratio just before her child's birth. Interestingly, the number of siblings and parental AFDC income in early adolescence seem to have a positive, rather than a negative, association with a woman's later self-sufficient income, although the coefficient is quite small. It also is important to note that neither race nor the age at which she gave birth are important predictors of a woman's self-sufficient income. While these factors are not central to our analysis, the results point to important family background characteristics that influence young women's abilities to be self-sufficient. These factors merit furtherexploration. To better understand how early economic supports boost later self-sufficiency, we add to the final model of Table 2 (the base model) variables measuring work, education, and family behaviors during the time the economic supports were received. We make these additions one at a time and then collectively. This allows us to first pinpoint which behaviors are the more important pathways through which economic supports are operating and then to assess the extent that these behaviors as a group represent all possible pathways. First, we add work experience to the original model. As Table 3 reveals, the number of years a woman works after the birth of her first 318 Journal of Marriage and the Family TABLE2. THEEFFECTOFEARLYECONOMIC SUPPORTS ONWOMEN'SLATERSELF-SUFFICIENT INCOME Predictors Earlyeconomic supports Aid to Families with DependentChildren Food stamps Child support Sharedhousing Financialsupportfrom relatives 1 -.228 (.178) -.413 (.453) 1.858*** (.283) 9.438*** (2.717) 1.144 (.970) Environmentalcontrols County unemploymentrate Countyunskilled wage rate Conditionsin early adolescence WhetherBlack Always rural Years poor AFDC income Numberof siblings Years with single parent Mother'seducation Low educationalexpectations 2 -.325 (.185) -.539 (.477) 1.700*** (.283) 9.245*** (2.746) .399 (.979) .524 (.992) -3.432* (1.398) -2.946 (4.029) 15.776 (9.312) -9.597 (9.808) -12.779** (4.835) .312** (.096) 3.905 (2.037) -6.533 (3.351) -1.085 (1.623) -17.932 (10.542) 13.525 (9.709) -6.931 (10.010) -19.249*** (5.594) .252* (.100) 4.429* (2.139) -6.343 (3.405) -.696 (1.619) -15.648 (10.717) State AFDC guarantee State sharepenalty Age at birth Year child was born 302 .1638 -.368 (.201) -.445 (.500) 1.680*** (.296) 10.054*** (2.797) -4.191** (1.298) -2.196 (3.789) Conditionsjust priorto first birth Individual'sparentalincome-to-needsratio Sample size (n) AdjustedR2 3 302 .2511 -9.366* (4.400) 1.888 (2.17) -2.021 (10.410) .857 (2.726) 1.253 (1.812) 302 .2634 Note: All income variables, including outcome, are measuredin hundredsof dollars ($100s). See Appendix for definitions of variables. *p < .05. **p <.01. ***p <.001. child is a large and highly significant predictor of her later self-sufficient income. Assuming the model is correctly specified, for every full-time equivalent year of work the young woman gets when her child is between ages 1 and 4, her selfsufficient income when this child is between ages 5 and 7 will increase annually by $2,500. Another factor that could influence a woman's later self-sufficiency is the amount of education she receives. To test how this behavior affects the relationship between economic support received and self-sufficiency, we add to the base model a variable representing years of education (see Table 3). Like work experience, education is a 319 Self-Sufficiencyfor YoungMothers TABLE3. THE EFFECTOFEARLYECONOMIC SUPPORTONWOMEN'SLATERSELF-SUFFICIENT INCOME, FORVARIOUSBEHAVIORS CONTROLLING Predictors Earlyeconomic support Aid to Families with Dependent Children(AFDC) Food stamps Child support Sharedhousing Financialsupportfrom relatives Base Model With Work With Education -.368 (.201) -.445 (.500) 1.680*** (.296) 10.054*** (2.797) .524 (.992) -.118 (.188) -.140 (.461) 1.809*** (.272) 10.189*** (2.568) 1.047 (.914) -.296 (.199) -.265 (.495) 1.583*** (.292) 9.092** (2.766) -.226 (1.002) Education,work, and family behaviors Workexperience 25.287*** (3.466) Education 9.173** (2.799) With With Additional With All Children Behaviors Marriage -.530* (.210) -.664 (.507) 1.643*** (.294) 7.940** (2.948) .465 (.986) - Additionalchildren Sample size (n) AdjustedR2 302 .2634 302 .3792 302 .2881 - -6.366* (2.956) Marriage -.261 (.188) -.285 (.465) 1.577*** (.275) 5.341* (2.690) -.098 (.926) 302 .2728 - -36.804*** (5.446) 302 .3647 -.105 (.185) -.059 (.441) 1.604*** (.256) 4.532 (2.619) -.188 (.878) 21.318*** (3.313) 8.156** (2.505) -4.220 (2.617) -26.398*** (5.234) 302 .4618 Note: Other control variables included in these models are county unemployment rate, county unskilled wage rate, whether Black, always rural,years poor, AFDC income, numberof siblings, years with single parent,mother's education, low educationalexpectations,individual's parentalincome-to-needsratio, state AFDC guarantee,state share penalty, age at birth,and year child was born. All income variables,including outcome, are measuredin hundredsof dollars ($100s). *p<.05. **p<.01. ***p<.001. statistically significant predictor of the woman's later self-sufficient income. Assuming the model is correctly specified, every year of education a woman receives increases her self-sufficient income by $917 annually. Another potential pathway through which economic support may influence women's selfsufficient income is marriage. To test this hypothesis, we introduce to the base model a variable representing the number of years a young mother is married in the 4 years following the birth of her first child. The results of this regression reveal a statistically significant, negative association between marriage and women's self-sufficient income. This relationship may exist because having a husband early on deters a woman's human capital development and weakens her attachment to the labor market. If her husband is able to earn substantial wages to support the family, she will have little incentive to make sure she is able to do the same. As we will see, marriage has quite a different relationship to marriage-oriented, selfsufficient income. A final behavior that might account for the influence of early economic supports is having additional children after the first is born. In Table 3, column 5, we introduce into the base model the number of additional children the woman had before her first child was 5 years old. Again, this behavior was significantly related to later selfsufficiency. For each additional child a woman had before her first child was 5, her annual selfsufficient income was reduced by $3,600. Some, but not all, of these four variableswork, education, marriage, and additional children-actually do account for part of the relationship between economic support and women's later self-sufficiency. When work experience is added to the base model, we see that the coefficients on shared housing and child support remain essentially unchanged, indicating that work is not a pathway through which these forms of early 320 economic support exert their influence on later self-sufficiency. The effect of both child support and shared housing on the outcome appears to be independent of the women's work experience during the observed time. Unlike work experience, though, education does play a role in converting one form of early economic support-shared housing-to later selfsufficiency. The coefficient on shared housing becomes slightly smaller when education is added to the base model. This suggests that some of the effect of shared housing operates through a woman's education level. A young woman who shares housing with her relatives is likely to be in a better position to pursue education after she has a child. This may be why education accounts for some of the effect observed in the base model between shared housing and the outcome. Like work experience and education, marriage has little influence on the relationship between child support and women's subsequent self-sufficient income. Marriage does, however, reduce the size of the coefficient on shared housing by 20% when compared with the base model. Apparently shared housing tends to postpone marriage, with that postponement enhancing a young mother's own economic self-sufficiency. When the number of additional children is added to the base model, the coefficient on shared housing again is reduced, this time by almost 50%. Thus, the positive effect of shared housing on self-sufficient income also is reduced if the young woman decides to have additional children. Child support, however, remains a significant predictor, with the size of the coefficient nearly what it was in the base model. Finally, to see the overall effect that these behaviors in the realms of education, work, and family have on the relationship between economic support and self-sufficiency, we add the four behaviors simultaneously to the base model. Child support remains the only statistically significant, early economic support after we control for these pathways. This finding reveals that the positive association between child support and later self-sufficiency does not depend on any of these behaviors. It is an independent effect. This finding is particularly striking because child support is not a very common type of economic support for this population due to their unmarriedstatus at the time their child was born. In contrast, the association between shared housing and later selfsufficiency is completely explained by these behaviors. Shared housing is apparently associated Journal of Marriage and the Family with higher educational attainment, fewer years of marriage, and fewer additional births for this population of young mothers. Collectively, these factors account for a substantial share of the association between shared housing and economic self-sufficiency. The models in Table 3 are highly predictive of women's self-sufficient income in this sample of young mothers. The adjusted R2 for each model indicates that a considerable amount of the variance in this outcome is explained by the variables included here. In fact, the final model that includes all four behaviors accounts for 46% of the variance in the outcome. Predictive Analysis-Marriage-Oriented Self-Sufficient Income To examine comparable relationships for a marriage-oriented definition of self-sufficiency, we repeat the same step-by-step process using as the dependent variable our second measure of selfsufficiency, which adds income from the husband's labor as a component of self-sufficient income (see the Appendix for a definition of this outcome). In the initial model containing only the early economic supports as predictors (not presented here), AFDC, food stamps, child support, and shared housing all are significantly related to this marriage-orienteddefinition of selfsufficiency. AFDC, food stamps, and shared housing are negatively related, while child support is positively related. As the control variables for environmental conditions, early adolescent family conditions, and circumstances before the birth are added, only AFDC and child support remain significant predictors, as shown in Table 4, column 1. The results in this analysis are different from those found in Table 3. This was expected, given the different nature of a marriage-oriented definition of self-sufficient income. As noted earlier, the mean of this marriage-oriented measure of is considerably higher than self-sufficiency women's self-sufficient income. Thus, the amount of income from the husband's labor is largely driving the results of these models for women who marry after the birth of their first child. As the regressions in Table 4 reveal, child support remains a statistically significant predictor with this definition of self-sufficiency. Shared housing, on the other hand, shows no statistically significant relationship to this marriage-oriented definition. Instead, another form of economic 321 Self-Sufficiencyfor YoungMothers TABLE SUPPORT ONLATER MARRIAGE-ORIENTED 4. THEEFFECT OFEARLY ECONOMIC BEHAVIORS SELF-SUFFICIENT FORVARIOUS CONTROLLING INCOME, Predictors Focal predictors Aid to Families with Dependent Children(AFDC) Food stamps Child support Sharedhousing Financialsupportfrom relatives Base Model With Work With Education With Additional With All With Behaviors Children Marriage -1.947*** (.393) -1.599 (.977) 2.297*** (.578) -7.641 (5.466) -1.846 (1.939) -1.744*** (.395) -1.353 (.969) 2.401*** (.572) -7.532 (5.400) -1.423 (1.922) -1.788*** (.387) -1.204 (.962) 2.084*** (.568) -9.754 (5.376) -3.495 (1.948) -1.236** (.387) -.445 (.938) 2.489*** (.544) 3.490 (5.452) -1.536 (1.824) Education,work, and family behaviors Work experience -.733* (.362) .347 (.867) 2.329*** (.503) -3.716 (5.156) -2.959 (1.729) -57.180*** (10.957) 302 .3377 18.043** (6.520) 10.487* (4.930) 37.003*** (5.151) -57.483*** (10.303) 302 .4629 20.454** (7.288) 20.158*** (5.441) Education Years married -- 33.522*** (5.468) Additionalchildren Sample size (n) AdjustedR2 -1.780*** (.377) -1.352 (.936) 2.136*** (.553) -14.963** (5.411) -2.812 (1.864) 302 .2756 302 .2930 302 .3071 302 .3593 Note: Other control variables included in these models are county unemployment rate, county unskilled wage rate, whether Black, always rural,years poor, AFDC income, numberof siblings, years with single parent,mother's education, low educationalexpectations,individual's parentalincome-to-needsratio, state AFDC guarantee,state share penalty, age at birth, and year child was born. All income variables,including the dependentvariable, are measuredin hundredsof dollars ($100s). *p <.05. **p <.01. ***p <.001. support emerges as a significant predictor of later self-sufficiency-the amount of AFDC a young mother receives soon after the birth of her child. Early support in the form of AFDC is negatively related to later marriage-oriented, self-sufficient income. If the base model in Table 4 is correctly specified, every $100 of AFDC income received decreases marriage-oriented self-sufficiency by $190. It is interesting to speculate why AFDC income is an important predictor in this model that uses a marriage-oriented definition of selfsufficiency when it failed to predict a stricter representation of women's self-sufficient income. It could be that women who receive AFDC are less likely to marry or less likely to marry men who earn high incomes. In the stricter definition of women's self-sufficiency, AFDC income is not negatively related to the outcome. As we found with our first definition of selfsufficiency, the work, educational, and family be- haviors that a young woman engages in after the birth of her child are all significant predictors of marriage-oriented self-sufficiency. Years of fulltime work experience and years of education remain positively associated with the outcome. The size of the years of education coefficient is, in fact, twice as large as in the women's selfsufficiency models (Table 3), suggesting that, for this population, education is more important in predicting the young woman's likelihood of marrying or perhaps the income of her husband than it is in predicting her self-sufficiency independent of marriage. It is not surprising that the number of years of marriage is positively related to this marriage-oriented measure of self-sufficiency. Likewise, the association between having subsequent children and this outcome is negatively related, as expected. It is interesting to note that the relationships between early economic supports and the mar- 322 Journal of Marriage and the Family riage-orienteddefinition of self-sufficiency are not much alteredby the inclusion of these four behaviors.This suggeststhatthe long-termbenefits of child supportand the negativeassociation of AFDC derive from somethingother than the work, education, and family behaviors of the young women. It may be, for example, that the economicsupportsaffectthe pool of eligible marriage partners for this population of young women. DISCUSSION This analysis has tried to answer the question, "Whatare the associationsbetweenthe economic supportthat young, unmarriedwomen receive soon after the birth of their first child and their later self-sufficient income?" We have operationalized self-sufficiency in two ways: first, as women's self-sufficient income, then in a more marriage-oriented way that includes the husband'sincomeif she has married. When we comparethe resultsof these sets of analyses, child supportis the only type of economic supportthat is a consistent predictorof both types of self-sufficiency.Yet, we are unable to shed muchlight on the reasonfor its predictive strengthbecausenone of the behaviorswe investigatedappearsto be the pathwayfor this association. As we see in both Tables 3 and 4, controlling for education,work,andfamily behaviorhas little effect on the relationshipbetweenchild supportpaymentsand self-sufficientincome, regardless of how it is defined. These results suggest that the women who receive child supportmight be differentfrom the rest of the samplein ways not controlledfor in this analysis. Othereconomicsupportsvaryin theirrelationship to the two types of self-sufficient income. Sharinghousingappearsto be an importantstrategy for improvingeconomic prospectsof young mothers by enhancing their chances of getting more educationand of havingfewer childrenbefore their first child is 5 years old. But shared housingdoes not appearto be an importantsource of economic supportfor marriage-orientedselfsufficiency. Instead,AFDC income is important and negatively associated with this marriageorientedoutcome. These differencesarise becauseof fundamental differences between the two outcomes. The first-women's self-sufficientincome-is a strict representationof a woman's own economic efforts.As such, the model tries to understandhow the behaviorsof the young womanafterthe birth of her child influence this outcome. The other outcome introducesthe husband's income as a form of self-sufficient income for those women who are marriedwhen theirchild reachesschool age. Becausethe income of men tends to be substantiallyhigher than the income of women, the husband'sincometendsto dominatethe woman's income in this outcome. Thus, the models in Table4 end up predictingmore aboutmen's situations or women's abilities to marrymen with high incomethanthey do aboutthe women'sown economicconditions. These results emphasize the need to think carefullyaboutour definitionof self-sufficiency. This is importantfor developingempiricalmodels, as well as for informing public policy debates. Marriageis not a long-term solution for making all women self-sufficient, particularly amongthis populationwhose marriagesare especially frail (Furstenberg et al., 1987; Teti & Lamb, 1989). Consequently, we believe that women's self-sufficientincome is a more useful concept than a marriage-orienteddefinition of self-sufficiency for the development of public policy. If the womanmarriesbut then faces marital disruption,which occurs with a high rate of probabilitythese days, it is her incomeapartfrom her husbandthat she is likely to be able to maintain without taking on additional labor market obligations(althoughthere could be some added private transfersresultingfrom divorce that are uncountedin ourmeasure).If we areinterestedin makingyoung,unmarriedmothersable to support themselves without governmentassistance, our first definition of women's self-sufficiency is moreappropriate for empiricalresearch. The resultsof this analysisalso emphasizethat young, unmarriedmothersare a diverse population. Once they have their first birth, their life scriptis not written.Some are able to gain valuable workexperience,while othersare not. Some are able to finish theireducationor postponeadditional childbearing.Some get married,while others remainsingle. These behaviorshave significantramificationson theirlaterabilitiesto become self-sufficient. By including behaviors that influence a woman'sabilityto get a payingjob or to improve her human capital, this analysis shows that a woman'sbehaviorafterhavingan early pregnancy is important.While out-of-wedlockbirthsmay contribute to detrimental outcomes for young mothers,therearea numberof differentpathways 323 Self-Sufficiencyfor YoungMothers open to them to reduce these negative consequences.Gainingworkexperienceis an important way to increase a woman's ability to be selfsufficientlater, althoughthis is difficult without adequatechild care or other supportsfor singleparentfamilies. Pursuingeducationand delaying additionalchildbearingalso can have significant implicationsfor a young woman's ability to be self-sufficientin the future. Future Research Given the focus on young mothersin the recent debate over welfare reform,researchabout this population'seconomicfutureis quitetimely.The researchpresentedhere is not, however, definitive. Severaltopicsidentifiedby this projectmerit furtherattention.For one, additionalresearchis needed that conceptually explores appropriate definitions of self-sufficiency, particularlyfor women and for families headed by women. Futureempiricalstudiesshouldtry to operationalize this concept in a numberof differentways. Because the concepthas receivedso muchpublicattentionand is heraldedas a goal of public assistanceprogramsin the 21st century,more systematicresearchinto its natureis needed. Secondly, future research should explore in moredetailhow marriageinfluencesyoungmothers' long-termself-sufficiency.Withinthe frameworkdevelopedhere,husbands'incomescouldbe added as a type of early economic supportfor women who marriedaftertheirfirst child's birth. Thiswouldhelpto showhow thistypeof economic supportassists women interestedin building humancapitalthrougheducationor work experience or by delayinghavingadditionalchildren. Our study also documentsthe importanceof expandingtraditionalnotionsof economicsupport beyond cash assistance for this population of young,unmarriedmothers.Futureresearchshould incorporateexpandednotionsof supportin the developmentof analyticalmodels.It also shouldtry to cast more light on our understandingof how child supportincreasesthe self-sufficiencyof this populationof youngmothers. Finally,the PSIDdata,while makingthis analysis possible, also have somewhat limited its scope. We are able to measureself-sufficiencyno later than when these women's childrenare between 5 and7 yearsold becauseadditionalwaves of dataarenot yet available.Perhapsthe mothers' work,education,and family behaviorsthatare so importantin this analysisbecome even morepiv- otal the longertheirchildrenare in school. It will be importantto follow up on this researchwhen more data are available.In addition,some types of economic support-such as child care-and measuresof importantchild outcomesare not included in the PSID data base. Exploringanother dataset with this information,such as the National Longitudinal Study of Youth, could expand this researchconsiderably. Implications for Public Policy The lessons from this researchsuggest that it is importantto provide young, unmarriedmothers with adequate income when their children are first born.Since child supportplays an important role, public policy can assist young, unmarried mothersto become self-sufficientby establishing child supportordersand enforcingtheirpayment. Encouraging young mothers to share housing with others is anotherpotential strategyfor increasing their self-sufficiency. While AFDC income exerts a small, negative effect on selfsufficiency,it does so only in the contextof selfsufficiencythat includesrelianceon a husband's income. This analysisclearly shows that the steps that young, unmarriedmothers take following the birth of their first child are importantdeterminantsof theireconomic securityand thatof their children.Manyof these stepsaresubjectto public policy interventions.It seems especially importantto supportinterventionsthatmakeit possible for young mothers to gain work experience. safe and high qualitychild care, Transportation, and tax credits for work-related expenses are some of the public policies that can make work morefeasiblefor them.This researchalso reveals thatprogramsthathelp young,unmarriedmothers avoid additional unintended pregnancies and pursue their educations are critical. If we, as a society, are committed to helping these young mothers become truly self-sufficient, there is muchthatwe can do to assistthem. NOTE The authorswouldlike to thankMaryCorcoran,Sandra Danziger,andtwo anonymousreviewersfor theirhelpful comments.Partialfundingfor the analysiswas provided by the National Institutefor Child Health and HumanDevelopment, Grant 1-R01-HD28145-01A2, andby the RussellSage Foundation. 324 Journal of Marriage and the Family REFERENCES Ashenfelter, O., & Krueger, A. (1992). Estimates of the economic returns to schooling from a new sample of twins (WorkingPaperNo. 4143). Cambridge,MA: NationalBureauof EconomicResearch. Brooks-Gunn,J., & Furstenberg,F. (1987). Continuity and change in the context of poverty: Adolescent mothers and their children. In J. Gallagher & C. Ramey (Eds.), The malleability of children. Balti- more,MD:PaulBrookes. Bumpass, L., & McLanahan,S. (1989). Unmarried motherhood:Recenttrends,composition,and BlackWhitedifferences.Demography,26, 279-286. Caldwell,C., Antonucci,T., Jackson,J., Osofsky,J., & Wolford, M. (1995, March). The availability and use of social support in families with adolescent mothers. Paperpresentedat the Society for Researchon Child IN. Development.Indianapolis, Corcoran, M., Gordon, R., Laren, D., & Solon, G. (1992). The associationbetweenmen's economicstatus andtheirfamilyand communityof origins.Journal of Human Resources, 27, 575-601. Corcoran,M., & Adams,T. (1995). Familyand neighborhoodwelfaredependencyand son's laborsupply. Journal of Family and Economic Issues, 16, 239-264. Danziger,S. (1991).Education,earningsandpoverty.In D. Hornbeck& L. Salamons(Eds.),HumanCapital and America's Future. Baltimore, MD: Johns Hop- kinsUniversityPress. Duncan,G., & Hoffman,S. (1988). The use andeffects of welfare:A surveyof recentevidence.Social Services Review, 62, 238-257. Duncan,G., & Yeung,W. J. (1995). Extentand consequences of welfare dependence among America's children. Children and Youth Services Review, 17, 157-182. Ellwood, D., & Bane, M. J. (1985). The impact of AFDC on family structureand living arrangements. In R. Ehrenberg (Ed.), Research in labor economics. Goldscheider, F., & DaVanzo, J. (1985). Living arrangements and the transition to adulthood. Demography, 22, 545-563. Haveman, R., Wolfe, B., & Spaulding, J. (1991). Childhood events and circumstances influencing high school completion. Demography, 28, 133-157. Hayes, C. (1987). Risking the future: Adolescent sexuality, pregnancy, and childbearing: Vol 1. Washington, DC:NationalAcademyPress. Hill, M. (1990, May). Shared housing as a form of economic support for young, unmarried mothers. Paper presented at the annual meetings of the Population Associationof America,Toronto,Canada. Hill, M. (1992). The Panel Study of Income Dynamics: A user's guide. Newbury Park, CA: Sage. Hill, M., & Duncan,G. (1987). Parentalfamily income andthe socioeconomicattainmentof children.Social Science Research, 16, 39-73. Hoffman,S., Foster,E. M., & Furstenberg,F. (1993). Reevaluatingthe costs of teenage childbearing.De- mography, 30, 1-13. Hogan, D. (1986). Maternal influences on adolescent family formation. Current Perspectives on Aging and the Life Cycle, 2, 147-165. Hogan, D., Hao, L., & Parish, W. (1990). Race, kin networks, and assistance to mother-headed families. Social Forces, 68, 797-812. Kellam,S., Adams,R., Brown,C. H., & Ensminger,M. (1982). The long-termevolutionof the family structure of teenage and older mothers. Journal of Marriage and the Family, 44, 539-554. Knox, V., & Bane, M. J. (1992). The effects of child support payments on educational attainment (Working Paper No. H-92-4). Boston, MA: HarvardUniversity Press, John F. Kennedy School of Government. Mare,R., & Winship,C. (1991). Socioeconomicchange and decline of marriagefor Blacks and Whites. In C. Jencks & P. Peterson (Eds.), Urban underclass. Washington,DC:BrookingsInstitute. Greenwich,CT:JAIPress. Ellwood,D., & Crane,J. (1990). Familychangeamong Black Americans:What do we know? Journal of Martinson, K. (1995). Review of the research evidence on employment, education, and training and work programs for welfare recipients (Technical analysis paper). Washington, DC: U.S. Department of Health Furstenberg,F., Brooks-Gunn,J., & Morgan, S. P. Maynard, R. (1995). Teenage childbearing and welfare Economic Perspectives, 4, 65-84. (1987). Adolescent mothers in later life. Cambridge, England:CambridgeUniversityPress. F., & Crawford,A. (1978).Familysupport: Furstenberg, Helping teenage mothersto cope. Family Planning Perspectives, 10, 322-333. Furstenberg, F., & Brooks-Gunn, J. (1986). Teenage childbearing: Causes, consequences and remedies. In L. H. Aiken & D. Mechanic (Eds.), Applications of social science to clinical medicine and health policy (pp. 307-334). New Brunswick,NJ: RutgersUniversity Press. GeneralAccountingOffice. (1991). Mother-onlyfamilies. Washington,DC:GovernmentPrintingOffice. Geronimus,A., & Korenman,S. (1992). The socioeconomic consequencesof teen childbearingreconsidered. Quarterly Journal of Economics, 107, 1187- 1214. Geronimus, A., & Korenman, S. (1993). Maternal youth or family background?On the healthdisadvantages of infantswith teenagemothers.AmericanJournalof Epidemiology, 137, 213-225. andHumanServices. reform:Lessonsfroma decadeof demonstration and evaluation research. Children and Youth Services Review, 17, 309-332. McLanahan, S. (1985). Family structure and the reproduction of poverty. American Journal of Sociology, 90, 873-901. Miller, B., & Moore, K. (1990). Adolescent sexual behavior, pregnancy, and parenting: Research through the 1980s. Journal of Marriage and the Family, 52, 1025-1044. Moffit, R. (1992). Incentiveeffects of the U.S. welfare system. Journal of Economic Literature, 30, 1-61. Mott, F., & Marsiglio,W. (1985). Early childbearing andcompletionof high school.FamilyPlanningPerspectives, 17, 234-237. Murray, C. (1993, October 29). The coming White underclass. The Wall Street Journal, p. A14. Ortiz,V., & Fennelly,K. (1988). Earlychildbearingand employment among young Mexican-origin, Black, and White women. Social Science Quarterly, 69, 98795. Self-Sufficiencyfor YoungMothers 325 Presser, H. (1980). Sally's corer: Coping with unmarried motherhood. Journal of Social Issues, 36, 107129. Scheirer, M. (1983). Household structureamong welfare families: Correlates and consequences. Journal of Marriage and the Family, 45, 761-771. Sewell, W., Hauser, R., & Wolf, W. (1980). Sex, schooling, and occupational status. American Journal of Sociology, 86, 551-583. Solon, G. (1992). Intergenerational income mobility in the United States. American Economic Review, 82, 393-408. Teti, D., & Lamb, M. (1989). Socioeconomic and marital outcomes of adolescent marriage, adolescent childbirth, and their co-occurrence. Journal of Marriage and the Family, 51, 203-212. Trent, K., & Harlan, S. (1990). Household structure among teenage mothers in the United States. Social Science Quarterly, 71, 439-457. Trent, K., & Harlan, S. (1994). Teenage mothers in nuclear and extended households. Journal of Family Issues, 15, 309-337. Upchurch, D., & McCarthy, M. (1989). Adolescent childbearing and high school completion in the 1980s. Family Planning Perspectives, 21, 199-202. U.S. Department of Health and Human Services. (1995). Report to Congress on out-of-wedlock childbearing (DHHS Pub. No. PHS 95-1257). Hyattsville, MD: Public Health Service. Wilson, W. J. (1987). The truly disadvantaged: The inner city, the underclass and public policy. Chicago: University of Chicago Press. Wojkiewicz, R., McLanahan, S., & Garfinkel, I. (1990). The growth of families headed by women: 19501980. Demography, 27, 19-30. APPENDIX OFVARIABLES DEFINITIONS Outcomesof interest Women's self-sufficient income Marriage-orientedself-sufficient income Average annualincome from the woman's labor, child suport,and financial assistancefrom relatives when first child was aged 5-7. Data are reportedin hundredsof dollars. Average annualincome from the woman's labor, husband'slabor,child support,and financial assistancefrom relatives when first child was aged 5-7. Data are reportedin hundredsof dollars. Focal predictors-sources of economic support AFDC Average annualincome the household received from Aid to Families with DependentChildrenand other welfare programs,such as state general assistanceprograms,while the first child was between 1 and 4 years old. Data are reportedin hundredsof dollars. Food stamps Average annualvalue of food stampsreceived while the first child was 1-4 years old. Data are reportedin hundredsof dollars. Child support Average annualchild supportreceived while the first child was 1-4 years old. The definitionof child supportincome changes a numberof times duringthe PSID study. Sharedhousing Total numberof years the young mothersharedhousing with another family unit. Relative support Average annualincome received from relatives outside the family unit while the child was between 1 and 4 years old. Factorsthat focal predictorscould operatethrough-education, work, and family behaviors Workexperience Numberof full-time equivalentyears the motherworked in the paid labor force while her first child was between 1 and 4 years old. Education Amount of educationthe motherhad before the child was 5 years old. Assigned 14 years if there was missing data and the individualwas away at an educationalinstitution.Assigned 12 years if other missing data. Years married Numberof full calendaryears motherwas marriedwhile her first child was between 1 and 4 years old. Additionalchildren Additionalchildrenbornto the motherbefore the first child was 5 years old. Environmentalconditions Countyunemploymentrate Based on Bureauof LaborStatisticsmonthly informationon employment and earningswhen the first child was aged 5-7. Missing data are assigned the value of the nationalunemploymentrate for the uninstitutionallabor force. Journal of Marriage and the Family 326 APPENDIX DEFINITIONS OF VARIABLES (CONT.) Countyunskilled wage rate Average county wage rate for unskilled male labor when the first child was aged 5-7. After 1985, this variableis nongenderspecific. Additionalcontrols:Conditionsin early adolescence 1 = if Black, 0 = if otherrace. Measuredat age 12 and determinedfrom WhetherBlack data on head of household. 1 = lived in ruralarea (where largestcity in county was less than 25,000) Always rural when motherwas between 12 and 14 years old; 0 = if she did not. Numberof years motherwas poor when she was aged 12-14. Years poor AFDC income Average annualincome from AFDC while motherwas aged 12-14. Measuredin hundredsof dollars. Numberof siblings Numberof otherindividualsyoungerthan 18 years old living in the household when the motherwas between 12 and 14 years old. Years with single parent Numberof years when motherwas living with only one parentbetween ages 12 and 14. Mother'seducation Numberof years of educationthe woman's motherhad when she was age 12. If missing value, then assessed when she was aged 13 or 14. If no data available, assigned 12. Low educationalexpectation 1 = woman's parentsbelieve that at least 1 child may not finish high school; 0 = woman's parentsbelieve all childrenwould finish high school. Measuredonce when woman was between 7 and 12 years old because only asked in early years of the PSID. Missing datacoded 0. Additionalcontrols:Conditionsjust priorto the birthof the first child Parentalincome-to-needsratio Mother'sparentalincome-to-needsratio in the year before the first child was born. If the motherhad left her parent'shome, no data were available, and a dummyvariablewas createdand included in all regressionsto control for this situation. State AFDC guarantee Maximummonthly AFDC grantpaid in the state for a family of 4 during the year before the first child was born. Data compiled by RobertMoffit for 1968-89. State sharepenalty 1 = if the motherlived in a state with a reductionin AFDC benefits if she sharedhousing with othersjust before the birthof the child. 0 = no state penalty for this situation.Policy assessed in 1985 for continentalU.S. Mother's age when her first child was born. Age at first childbirth Year child born Last 2 digits of the year that the mothergave birthto her first child. For example, 1981 has a value of 81. Note: All dollar variableshave been convertedto 1990 values using the ConsumerPrice Index.