Nonmarital fertility refers to births occurring outside of formal marriage — to never-married women, and to women who are separated, divorced, or widowed. It is measured either as a period ratio (PNMFR: nonmarital births / all births in a calendar year) or, following Wu (2008), as a cohort ratio (CNMFR: completed nonmarital fertility / completed fertility for a birth cohort of women). The two measures can diverge substantially when cohort fertility trends differ from period trends.
The period nonmarital fertility ratio (PNMFR) from vital statistics is the standard official measure but carries two systematic biases:
Wu's cohort nonmarital fertility ratio (CNMFR = completed nonmarital fertility / completed fertility for women at survey) is the direct cohort analog, immune to these compositional artifacts. See Period Mortality for an analogous period/cohort tension in mortality measurement.
Using pooled June Current Population Survey (CPS) retrospective histories (1980, 1985, 1990, 1995), Wu documents a nearly monotonic rise across years of cohort experience:
| Birth Cohort | % with nonmarital birth by age 30 (All) | % White | % Black | % Hispanic |
|---|---|---|---|---|
| – | ||||
| – | ||||
| – |
The CNMFR (completed fertility basis) rose from to for all women and from to for white women between the pre-1925 and – cohorts — a threefold increase.
The proportion of women with at least one nonmarital birth roughly doubles between ages 20 and 30 for all cohorts of white and Hispanic women (+two-thirds for black women). Substantial nonmarital childbearing occurs in the 20s and early 30s across all cohorts and groups; treating nonmarital fertility as synonymous with teen childbearing mismeasures both the magnitude and the social mechanisms involved.
White nonmarital births are more likely than black or Hispanic nonmarital births to follow marital separation or divorce (rather than occurring to never-married women). Across all periods:
This compositional difference implies different social mechanisms: white nonmarital fertility is more frequently a post-marital dissolution phenomenon (cohabitation, repartnering); black nonmarital fertility is more frequently a pre-marriage phenomenon (never-married women). Policies calibrated to one mechanism are poorly targeted to the other.
U.S. total fertility has remained near replacement levels ( children per woman) in part because nonmarital births sustain aggregate fertility even as marital fertility has fallen. Countries with very low nonmarital fertility (Japan, southern Europe) tend to have very low total fertility rates. Rindfuss, Guzzo, and Morgan (2003) document a positive cross-national association between nonmarital fertility and total fertility rates across 22 industrialized nations. This creates a policy paradox: interventions aimed at reducing nonmarital fertility may carry the unintended consequence of reducing total fertility to below-replacement levels.
The same cross-national pattern appears in the Engelhardt and Prskawetz (2004) analysis of 21 Organisation for Economic Co-operation and Development (OECD) countries: Mediterranean countries with low nonmarital fertility (Italy, Spain, Greece) combine that with the lowest total fertility rate (TFR) in the sample. The mechanism is distinct from the Nordic pattern — Mediterranean countries condition childbearing strongly on stable marriage and male breadwinner economic position. When male labor market conditions deteriorate (rising male unemployment, declining relative wages), marriage and fertility decline together. The result is a "neither-work-nor-children" trap rather than a "children instead of work" specialization. See Fertility-Employment Correlation.
One of the most counterintuitive facts in U.S. family demography is that non-marital fertility rose sharply from the late 1960s through the late 1980s despite dramatic improvements in contraceptive technology. The standard expectation — better contraception → fewer unplanned pregnancies → fewer non-marital births — was not borne out.
Akerlof, Yellen, and Katz (1996) explain the paradox through a change in the shotgun marriage norm. Before the oral contraceptive pill became widely available to unmarried women (via Eisenstadt v. Baird, 1972), pregnancy outside marriage carried strong social and implicit contractual pressure on men to "make an honest woman" of their partner — the shotgun marriage. This norm operated because women's expected fertility from premarital sex was high and shared equally between partners: men had roughly symmetrical reason to avoid premarital sex (or to commit quickly on pregnancy), and women could credibly demand marriage as a condition of the relationship.
The pill disrupted this equilibrium. Once reliable contraception was available, a woman who became pregnant despite having access to contraception was seen as having "chosen" pregnancy — and could no longer credibly invoke the old social norm to demand marriage. The key result: women lost their ability to extract marriage commitments on pregnancy even as abortion remained available as an alternative. The net effect was a decline in the shotgun marriage rate that outweighed the decline in unplanned pregnancies, producing a rise in non-marital births from an equilibrium with better contraception.
This mechanism is relevant to interpreting the Wu (2008) cohort data: the steep portion of the rise in CNMFR for white and Hispanic women (roughly the 1940s–1965 birth cohorts) coincides with the period when the pill was diffusing and the shotgun marriage norm was collapsing — not with a period of worsening contraception or declining male earnings. See also Marriage Market for the production-to-consumption-complementarity reframing of modern marriage that contextualizes why this institutional norm did not reform itself.
Cohabitation was nearly absent in U.S. data before 1970 and has grown continuously since. By 2002, 50% of women aged 15–44 had ever cohabited; 59% of early-2000s first marriages were preceded by cohabitation with the future spouse (75% for remarriages). Pre-marital cohabitation is historically associated with higher subsequent divorce rates, though this reflects selection (higher-risk couples cohabit rather than marry) rather than a causal effect of cohabitation per se. Cohabitation may account for up to one year of the observed rise in median age at first marriage.
From an international comparative perspective, cohabitation is not a U.S.-specific phenomenon but shows divergent trajectories: Sweden has institutionalized cohabitation as a near-marriage substitute (non-marital births 55.4%, cohabitation rate 11.7%), while Italy maintains very low cohabitation alongside very low divorce and very low total fertility — illustrating that the relationship between cohabitation, marriage, and fertility depends heavily on the institutional and welfare context. The U.S. trajectory is distinctive in combining relatively high cohabitation rates with persistently high (by developed-country standards) marriage, remarriage, and total fertility rates.
Wu (1996) asks not what the trend looks like but who is at risk. Using discrete-time hazard models on the National Longitudinal Survey of Youth (NLSY) (female respondents, birth cohorts –), Wu isolates three independent predictors of a first premarital birth:
The three effects are additive and independent, meaning income transfers alone (raising level) would not eliminate the instability penalty, and family-stability interventions alone would not address the economic pathways.
Wu uses life table methods to estimate the cumulative probability of at least one nonmarital birth by ages 20, 25, and 30, handling right-censoring (women in younger cohorts not yet observed to age 30 at survey). The CPS pooling provides sufficiently large minority-group samples despite no oversampling.
The definition of a nonmarital birth is operationalized as births occurring before marriage, to never-married women, or after marital separation — rather than the official definition of births to formally unmarried women. This matters because formal divorce can lag separation by months to years; births between separation and divorce that are conceived after separation are functionally nonmarital. Wu shows that of births occurring between separation and divorce happen months post-separation (plausibly nonmarital conceptions).
Wu's cohort trends are the long-run demographic backdrop for the Autor, Dorn, and Hanson (ADH) 2019 causal story: manufacturing decline reduces male marriageability more women remain unmarried more children born outside marriage. The Wu data documents the 50-year trend that ADH explains a portion of causally (the portion attributable to trade shocks in specific commuting zones). See Marriage Market.
Angrist (2001) provides a third, historically earlier causal link in the same chain. Using U.S. immigration quota legislation as an instrument for immigrant sex ratios, he shows that higher male-to-female ratios among early 20th-century immigrant communities causally raised female marriage rates among second-generation Americans — the direct inverse of the marriage-rate decline documented by ADH and the nonmarital fertility rise documented by Wu. The convergence of evidence across three different identification strategies (immigration legislation, trade shocks, and cohort trends) and three different time periods (–, –, and – birth cohorts) establishes the sex-ratio marriage-rate link as one of the most robustly documented relationships in family economics. See Angrist 2001 — How Do Sex Ratios Affect Marriage and Labor Markets Evidence from Americas Second Generation.
Fraction of single-parent households is the single strongest correlate of low intergenerational upward mobility across U.S. commuting zones (; Chetty et al. 2014c). Wu's cohort data traces the underlying trend driving this correlation — the rise is not a one-decade artifact but a 50-year structural shift across successive birth cohorts. See Intergenerational Income Mobility.
The Fragile Families and Child Wellbeing Study (McLanahan 2009) identifies the two proximate mechanisms through which non-marital childbearing reproduces disadvantage. Despite high optimism at birth ( of cohabiting mothers expect to marry), by year 5 only have married and only of unmarried parents are living together.
Partnership instability: By the child's third birthday, two-thirds of unmarried mothers have experienced partnership change, have experienced , and have experienced (vs. of married mothers experiencing any change). Each change increases maternal stress and reduces parenting quality, especially for mothers with high school education. Critically, children in stably cohabiting households are statistically indistinguishable from children in stably married households in aggression and anxiety/depression; the harm comes from instability, not from the absence of a marriage certificate. Half of the family-structure/behavior-problem association is mediated through maternal stress and poor parenting.
Multi-partnered fertility (MPF): – of unmarried parents have children with different partners (vs. – of married parents). Fathers' MPF — more than mothers' — destabilizes current unions: it drains paternal resources, generates jealousy, and reduces co-parental cooperation. MPF thus spreads fathers' time and money across multiple households, reducing effective paternal investment in any given child. See Multi-Partnered Fertility.
The nonmarital fertility rise matters not only as a demographic trend but because father absence in nonmarital households has documented causal effects on child outcomes. McLanahan, Tach, and Schneider (2013) review 47 studies using rigorous causal designs and find that father absence causally reduces high school graduation ( studies significant), increases externalizing behavior ( studies significant), and increases substance use ( studies significant) and adult mental health problems ( studies significant) — even after correcting for selection into nonmarital birth and father departure. Test score and earnings effects are weaker and less consistent. The causal mechanism operates primarily through behavioral/noncognitive pathways: externalizing behavior and school disengagement, rather than cognitive ability deficits. Boys are more affected than girls, consistent with the family disadvantage literature (Autor et al. 2019). See Causal Effects of Father Absence and Gender Gap in Educational Outcomes.
Children born outside of marriage and their mothers are disproportionately likely to live in poverty. Wu frames this as motivation for cohort-level trend data: without knowing how the population of women experiencing nonmarital fertility has changed across cohorts (in composition, age distribution, and type of union dissolution), welfare policy cannot be accurately targeted. The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA)'s 1996 restructuring of welfare was premised partly on assumptions about nonmarital fertility that Wu's cohort methods help evaluate more precisely.