Marriage is becoming less a foundation for economic security than a signal of having already achieved it. As education, housing and stable employment become more unequally distributed, so do the conditions under which people form lasting families. The United States shows how that divide can carry advantage across generations; Europe shows that its consequences depend on institutions; Italy is the test of whether the American pattern spreads.
In 1970, Americans with a university degree were slightly less likely to be married than high-school graduates: 79% against 81% of adults aged 25 and over. By 2024 they were fourteen percentage points more likely, 64% against 50%. Marriage had not become more popular among the educated. It had retreated faster among everyone else.
That reversal is the subject of this Dispatch. The institution that once marked the beginning of adult life is becoming a reward for having already secured one. The question is no longer only why fewer people marry. It is why the people best equipped to live without marriage are increasingly the ones who still have it.
The argument separates three things that are often run together. Marriage is who enters the legal institution. Union stability is who sustains a lasting partnership, married or not. Childhood stability is the family setting in which children grow up. It also separates two processes: the stratification of family formation, meaning how class shapes who forms and keeps unions, and the stratification of consequences, meaning how much a less stable family costs the children in it. The claim is that the first process is best explained by the “capstone” logic of modern marriage, and that the second depends largely on institutions. The claim would be weakened if less-educated Europeans proved to be substituting stable cohabitation for marriage as fully as graduates do, or if the convergence now visible among young Italian cohorts stalled.
The Vanishing Institution
Across the European Union the crude marriage rate more than halved between 1964 and 2024, from 8.0 to 3.9 per 1,000 inhabitants, while the crude divorce rate doubled, from 0.8 to 1.6. In 2024, 41.1% of births in the EU took place outside marriage, against 17.7% in 1993, the first year Eurostat publishes the aggregate; because national coverage of the series has changed, the comparison indicates direction and scale rather than an exact increase. In six member states births outside marriage were the majority: Bulgaria (61.9%), France (59.7%), Portugal (59.2%), Sweden (57.1%), Slovenia (55.5%) and Estonia (50.7%). Italy celebrated 173,272 marriages in 2024, six in ten of them civil, and provisional figures for the first nine months of 2025 show a further fall of 5.9%.
The usual explanation is emancipation. Contraception, women’s education and employment, divorce reform and secularisation turned marriage from an obligation into an option. Demographers group these changes under the Second Demographic Transition, a long-run shift towards autonomy and self-realisation. The explanation is correct but incomplete, because the transition changed not only how often people marry but what marriage means.
In 2004 the sociologist Andrew Cherlin described the “deinstitutionalisation” of American marriage: the weakening of the norms that once told people whom to marry, when, and in what order relative to sex, cohabitation and children. He drew a distinction that frames everything that follows. Marriage used to be a marker of conformity; it has become a marker of prestige. In the older cornerstone model, people married early and built their adult lives together. In the capstone model, marriage comes last, as the finishing stone on a structure already standing: a degree, a steady job, a home, savings, a partner tested over time.
A capstone depends on what is beneath it. If marriage is the reward for a secure life, then marrying depends on having been able to build one, and the materials are not distributed evenly.
The New Class Divide
The evidence for the second transformation is strongest in the United States.
| Year | Less than high school | High-school graduate | Some college | Bachelor’s or more |
|---|---|---|---|---|
| 1970 | 69 | 81 | 78 | 79 |
| 1990 | 57 | 67 | 67 | 69 |
| 2010 | 48 | 55 | 57 | 66 |
| 2024 | 48 | 50 | 54 | 64 |
The table must be read with care. Graduates did not marry more; their married share fell by fifteen points. Everyone else’s fell by between twenty-one and thirty-one. The gap opened because the decline was uneven, and the near-identical figures for 2019 and 2024 suggest it has stopped widening.
The divide began with stability, not with entry. Among Americans born between 1957 and 1964, almost everyone eventually married: 89% of graduates and 87% of high-school graduates by age 46. But by 46, 49% of the high-school graduates who had married were divorced, against 30% of the graduates (Lundberg, Pollak and Stearns, Journal of Economic Perspectives, 2016).
Only later did the gap reach entry itself. In 2023, never-married Americans with a master’s degree entered a first marriage at about 84 per 1,000 a year; those without a high-school diploma at about 27, less than a third of that rate (National Center for Family & Marriage Research, 2025). Exit follows the same line. Among Americans who have ever married, 25% of graduates have been through a divorce, against 37% of those with a high-school education or less and 41% of those with some college (Pew Research Center, 2025). That last measure is cumulative, not an annual risk, and is sensitive to age and marital history; it describes accumulated experience, not this year’s behaviour.
Entry and exit together explain an apparent paradox. The American refined divorce rate, divorces per 1,000 married women, peaked at about 22.6 around 1980 and fell to 14.4 in 2023. Pew attributes the fall partly to the married population shifting towards the better educated, who divorce less. Marriage looks more stable in part because those most at risk of divorce increasingly do not marry at all.
Cherlin has since drawn the same conclusion about his own thesis. In a 2020 reassessment he found that deinstitutionalisation had occurred mainly among Americans without a college degree, while among graduates marriage “still plays a central role”. The traditional institution has survived at the top. Below it, family life has moved towards what he calls the “marriage-go-round”: cohabitation, separation and new partners in sequence.
The Price of Being Ready
The first explanation is economic. Cherlin’s Labor’s Love Lost (2014) links the decline of working-class marriage to the loss of the industrial jobs that once let a man without a degree support a family. The strongest causal evidence comes from David Autor, David Dorn and Gordon Hanson (AER: Insights, 2019). They compared American labour markets by their exposure to Chinese import competition. Where the shock lowered young men’s earnings relative to women’s, marriage and fertility fell, and the share of children born to young unmarried mothers and raised in poor single-parent households rose.
Money alone, however, does not explain the pattern, and a pair of natural experiments shows why. Melissa Kearney and Riley Wilson (Review of Economics and Statistics, 2018) studied the fracking boom of the 2000s, which raised wages and jobs for men without degrees. Births rose, inside and outside marriage, but marriage rates did not. They contrasted this with the Appalachian coal boom of the 1970s and 1980s, a comparable shock to men’s earnings, which raised marriage rates and marital births but not non-marital births. The shock was similar; the response was not. The authors read the contrast as suggesting an interaction between economic forces and social context. It is consistent with a change in the meaning of marriage between the two booms: by the 2000s, higher income was enough to start a family but no longer, on its own, enough to make marriage the expected next step. The comparison is suggestive rather than conclusive, since the two booms also differ in period, region and population.
What changed is described most clearly in fieldwork. Over five years, Kathryn Edin and Maria Kefalas interviewed and observed 162 low-income single mothers in eight neighbourhoods of Philadelphia and Camden (Promises I Can Keep, 2005). These women did not reject marriage. They held it in high regard, and that was part of the difficulty. They believed it should come only after financial independence and a relationship proven over time, and they feared divorce more than single motherhood. Motherhood was a commitment they felt able to keep. Marriage was one they felt they could not yet make.
This is the capstone model seen from below. Economists describe it from above. Shelly Lundberg and colleagues argue that for graduates, marriage works as a commitment device supporting long, costly, jointly planned investment in a small number of children. Where the future is predictable, a long-term contract is valuable. Where it is not, the same contract looks less like insurance and more like a risk.
The word in the title needs a precise reading. In economics a luxury good is one whose demand rises more than proportionally with income, and this Dispatch does not measure that elasticity. What the evidence does show is closer to a status good: marriage has become a public signal of security already achieved, and that signal is unequally distributed. Here the strongest objection to the thesis turns out to support it. Britain’s Institute for Fiscal Studies found that differences in child outcomes between married and cohabiting parents are largely explained by who the parents are, especially their education, and saw no strong case for promoting marriage to improve children’s outcomes (Crawford, Goodman and Greaves, 2013). If marriage mostly reflects security rather than producing it, the certificate may change little in itself. What it signals is the point.
What Children Inherit
The three levels named at the outset now matter. A decline in marriage is not by itself a decline in two-parent families: across most OECD countries, the large majority of children still live with two parents, including where marriage has become uncommon (OECD Family Database). The question is not whether children’s parents are married but whether their families are stable, and here the American gradient is steep.
| Mother has a four-year degree | Mother has high school or some college | |
|---|---|---|
| Births outside marriage, 1980 → 2013 | 5% → 11% | 24% → 58% (high school) |
| Children living with two married parents, 1980 → 2019 | 90% → 84% | 83% → 60% |
Exhibit 2 measures married parents, not two parents, so it overstates the loss of two-parent homes to the extent that cohabiting couples fill the gap. In the United States that offset is small: the Census Bureau counted about 3.2 million children living with cohabiting parents in 2023, roughly one child in twenty-three. Unlike much of Europe, American cohabitation has not replaced marriage as a setting for raising children. Sara McLanahan called the result “diverging destinies” in her 2004 presidential address to the Population Association of America: the same demographic transition brought the children of the most educated mothers more parental time and money, and the children of the least educated relative losses. Kearney’s word privilege means an advantage received without being earned, now distributed by class.
Most of these figures are associations, not proof that marriage itself improves children’s lives. The best causal review, by McLanahan, Laura Tach and Daniel Schneider (Annual Review of Sociology, 2013), examined studies using fixed effects, natural experiments and propensity-score matching. Negative effects of father absence persist under these designs but are smaller than simple comparisons suggest, and are clearest for high-school completion, social and emotional adjustment, and adult mental health. What this research supports is the importance of stable parental presence, not of the legal form of marriage.
Two further strands of evidence bear on different questions and should be kept apart. The first concerns adults’ well-being, not family stability. Brienna Perelli-Harris and colleagues (Demography, 2019) compared the United Kingdom, Australia, Germany and Norway and found that, once selection and relationship quality are accounted for, the life-satisfaction advantage of married over cohabiting adults disappears. That result says marriage does not by itself make partners happier; it says nothing directly about how long unions last or how children fare. The second strand concerns stability, and here cohabitation and marriage still differ. Harmonised survey data from sixteen European countries, analysed by authors who openly favour marriage, found children born to cohabiting parents about 90% more likely to see them separate by age twelve (Wilcox and DeRose, Brookings, 2017).
The idea of a hereditary family should therefore be stated as a plausible mechanism, not a proven result. The conditions that make a lasting partnership look reasonable (a predictable income, a home, the ability to plan decades ahead) are themselves unequally passed down. Children raised in stable homes do better at school, and graduates tend to partner with graduates. Tracing the full transmission of family behaviour from parents to children requires studies that follow both generations, and that is where the evidence is thinnest.
The European Test
Europe is where the thesis meets its hardest test, and where it has to be refined.
Matthijs Kalmijn’s comparison of 25 European countries (Demography, 2013) found that the marriage gradient is not universal. Where gender roles are traditional, better-educated women are less likely to be married; in gender-egalitarian countries the relationship reverses. For men, education has no effect in traditional countries and becomes positive as equality grows, and the effect is stronger where economic inequality between education groups is large. The American pattern is therefore not a law of modernity. Kalmijn’s results associate it with a particular combination: gender equality together with large economic inequality between education groups. That is a cross-sectional pattern across countries, not a demonstrated causal mechanism.
A comparison of the United States with fourteen European countries points the same way (Perelli-Harris and Lyons-Amos, 2016). Educated people everywhere marry later, but strong educational divergence in marriage and separation trajectories is clear in the US and weak elsewhere, and national context explains more than education. On divorce, by contrast, Europe is converging. A meta-analysis of European longitudinal studies (Matysiak, Styrc and Vignoli, Population Studies, 2014) found that the old positive gradient, in which the educated divorced more, has weakened and in some countries reversed, mainly because of women’s empowerment rather than changes in divorce law.
| Context | Marriage by education | Divorce by education | Representative finding |
|---|---|---|---|
| United States | Strongly positive; widened after 1980 | Negative: graduates divorce least | Pew; NCFMR; Lundberg et al. 2016 |
| United Kingdom | Positive and widening | Negative | IFS Deaton Review, Families and inequalities (2022) |
| Nordic countries | Positive: marriage fell most among the least educated | Negative | Jalovaara and Andersson, Demographic Research (2023), Finland |
| Gender-traditional Europe (historically) | Negative for women | Positive: the educated divorced more | Kalmijn 2013; Härkönen and Dronkers 2006 |
| Italy, youngest cohorts | Not yet established | Vanishing among women, reversed among men | Bastianelli, Guetto and Vignoli 2024 |
The Nordic case refines the thesis rather than refuting it. Finnish registry data for 1989–2019 show marriage falling in every education group but most steeply among those with only basic schooling, while the advantage of the educated stayed broadly stable. The stratification of family formation persists. What differs is the stratification of consequences: Juho Härkönen’s comparison of single motherhood across rich countries (2017) finds that the link between family structure and child poverty depends on how much the welfare state offsets it. The Nordic welfare state does not make families equal. It reduces how much their inequality is passed on to children.
Italy, the Latecomer
In 1973 Italy recorded about 418,000 weddings, almost all of them first marriages and 95.9% religious. Women born in 1933 turned forty that year, and 879 in every 1,000 had married.
| Born 1933 | Born 1958 | Born 1983 |
|---|---|---|
| 879 | 870 | 578 |
For decades Italy was Europe’s clearest counter-example. In a society with traditional gender roles, marriage cost educated women the most, and they were less likely to marry. Divorce, introduced in 1970, was more common among the educated, the reverse of the American pattern. In an eight-country study, Italy was the only country where childbearing within cohabitation was not concentrated among the less educated (Perelli-Harris et al., Population and Development Review, 2010). The regional map also reverses America’s: the poorer South marries more, younger and in church, while civil weddings, separations and divorces concentrate in the North.
That exception is eroding. Using ISTAT’s 2016 family survey, Arnstein Aassve, Letizia Mencarini, Elena Pirani and Daniele Vignoli find cohabitation, separation and non-marital childbearing moving steadily in the direction the Second Demographic Transition predicts, with the South following more slowly (“The Last Bastion is Falling”, Population and Development Review, 2024). Elena Bastianelli, Raffaele Guetto and Vignoli (Research in Social Stratification and Mobility, 2024) find that among younger cohorts the educational and class gradient in marital breakdown has vanished for women and reversed for men. Among cohabiting couples there is no clear socioeconomic gradient, but cohabiting men without work face a much higher risk of breaking up.
This is evidence of convergence in union dissolution. It is not yet evidence that educated Italians have gained an American-style advantage in entering marriage or in sustaining unions overall. The conditions that would produce such an advantage are nonetheless unusually present. The share of 18–34-year-olds living with their parents rose from 59.7% in 2002 to 67.4% in 2022, above 75% in Campania and Puglia. Precarious contracts and housing costs delay independence, and a capstone cannot be placed until the life beneath it is built.
The fairest conclusion is a prediction rather than a verdict. Italy may be developing the conditions for an American-style class divide in family stability, but the outcome is not settled. One serious alternative view holds that the instability is transitional: Gøsta Esping-Andersen and Francesco Billari (Population and Development Review, 2015) argue that as gender equality matures, families restabilise and class gradients may weaken. Falling divorce across the rich world is consistent with that view, but a fall produced by selection, with the most at-risk couples no longer marrying, is a retreat of the institution rather than a return of the family.
The test is specific. If the capstone logic is right, then among the Italian cohorts now in their twenties and thirties, education and economic security will become steadily stronger predictors of forming and keeping a stable union, married or cohabiting. If they do not, the Italian exception will have survived its last test. Either way, Italy combines the conditions that drive the stratification of formation (late, expensive housing and a precarious labour market for the young) with limited family support. On the Nordic evidence, the size of the second stratification, the one that falls on children, is a matter of policy rather than of demography.
Marriage has not disappeared. Its social distribution has changed. It began as a universal stage of life, became an individual choice, and is increasingly a status available to those who already meet an informal requirement: being settled.
That is the paradox the last half-century leaves behind. The freedom not to marry has become universal. The conditions for building a family that lasts, married or not, have not. A society can open every option to everyone and still find that only some of its members are in a position to choose the stable one, and that their children inherit the position as well as the choice.
Notes on the data. Exhibit 1 measures a stock (adults currently married). First-marriage rates are annual flows per 1,000 never-married adults aged 18+ (NCFMR, FP-25-01); the Pew divorce shares are cumulative experience among the ever-married. Refined divorce rates are per 1,000 married women. EU crude rates: Eurostat (demo_nind, demo_ndivind). ISTAT’s mean age at first marriage covers ages 16–49 and is lower than Eurostat’s figure for Italy. Wilcox and DeRose are attributed as authors who advocate for marriage. Causal claims rest only on Autor, Dorn and Hanson; Kearney and Wilson; and McLanahan, Tach and Schneider. All other relationships reported here are associations. “Luxury” in the title is used in its editorial, not its technical economic, sense.
Sources
- Pew Research Center, analysis of 1970–2000 decennial census and 2010, 2019 and 2024 American Community Survey (IPUMS), March 2026: adults currently married by education; children living with two married parents.
- Pew Research Center, “8 facts about divorce in the United States”, 16 October 2025.
- National Center for Family & Marriage Research, Bowling Green State University, First Marriage Rate, 2023 (FP-25-01), 2025; Refined Divorce Rate, 2024.
- Shelly Lundberg, Robert A. Pollak and Jenna Stearns, “Family Inequality: Diverging Patterns in Marriage, Cohabitation, and Childbearing”, Journal of Economic Perspectives 30(2), 2016.
- Andrew J. Cherlin, “The Deinstitutionalization of American Marriage”, Journal of Marriage and Family 66(4), 2004; “Degrees of Change: An Assessment of the Deinstitutionalization of Marriage Thesis”, Journal of Marriage and Family 82(1), 2020; Labor’s Love Lost, Russell Sage Foundation, 2014.
- David Autor, David Dorn and Gordon Hanson, “When Work Disappears: Manufacturing Decline and the Falling Marriage Market Value of Young Men”, AER: Insights 1(2), 2019.
- Melissa S. Kearney and Riley Wilson, “Male Earnings, Marriageable Men, and Nonmarital Fertility: Evidence from the Fracking Boom”, Review of Economics and Statistics 100(4), 2018.
- Kathryn Edin and Maria Kefalas, Promises I Can Keep: Why Poor Women Put Motherhood Before Marriage, University of California Press, 2005.
- Melissa S. Kearney, The Two-Parent Privilege: How Americans Stopped Getting Married and Started Falling Behind, University of Chicago Press, 2023.
- Claire Crawford, Alissa Goodman and Ellen Greaves, Cohabitation, Marriage, Relationship Stability and Child Outcomes, Institute for Fiscal Studies, 2013.
- Sara McLanahan, “Diverging Destinies: How Children Are Faring Under the Second Demographic Transition”, Demography 41(4), 2004; Sara McLanahan, Laura Tach and Daniel Schneider, “The Causal Effects of Father Absence”, Annual Review of Sociology 39, 2013.
- US Census Bureau, America’s Families and Living Arrangements: 2023; OECD Family Database, SF1.2.
- Brienna Perelli-Harris, Stefanie Hoherz, Trude Lappegård and Ann Evans, “Mind the ‘Happiness’ Gap”, Demography 56(4), 2019; W. Bradford Wilcox and Laurie DeRose, “In Europe, Cohabitation Is Stable… Right?”, Brookings, 2017.
- Matthijs Kalmijn, “The Educational Gradient in Marriage: A Comparison of 25 European Countries”, Demography 50(4), 2013; Brienna Perelli-Harris and Mark Lyons-Amos, partnership trajectories in the US and 14 European countries, 2016.
- Anna Matysiak, Marta Styrc and Daniele Vignoli, “The Educational Gradient in Marital Disruption: A Meta-analysis of European Research Findings”, Population Studies 68(2), 2014; Juho Härkönen and Jaap Dronkers, European Sociological Review 22(5), 2006.
- Institute for Fiscal Studies, Deaton Review of Inequalities, Families and inequalities, 2022; Marika Jalovaara and Gunnar Andersson, Demographic Research 48, 2023; Juho Härkönen, single motherhood and child poverty across rich countries, 2017.
- ISTAT, Matrimoni, unioni civili, separazioni e divorzi – Anno 2024, 19 January 2026; ISTAT, Rapporto annuale 2025, ch. 3.
- Arnstein Aassve, Letizia Mencarini, Elena Pirani and Daniele Vignoli, “The Last Bastion is Falling”, Population and Development Review 50(4), 2024; Elena Bastianelli, Raffaele Guetto and Daniele Vignoli, “The changing educational and social class gradients in union dissolution”, Research in Social Stratification and Mobility 92, 2024; Brienna Perelli-Harris et al., Population and Development Review, 2010.
- Gøsta Esping-Andersen and Francesco C. Billari, “Re-theorizing Family Demographics”, Population and Development Review 41(1), 2015.
- Eurostat, demo_nind, demo_ndivind and demo_find; 2024 release, March 2026.
- Context
- The continued fall of marriage across the rich world: ISTAT's 2024 figures (173,272 marriages, a further 5.9% fall in early 2025), Eurostat's 2024 data and Pew's March 2026 analysis of fifty years of American family change.
- Causes
- The shift from 'cornerstone' to 'capstone' marriage; the falling economic position of men without degrees; unequal access to stable work and housing; welfare institutions that set how much family instability costs children.
- Actors
- Andrew Cherlin, Sara McLanahan, Melissa Kearney, Kathryn Edin and Maria Kefalas, David Autor, Shelly Lundberg; Matthijs Kalmijn, Brienna Perelli-Harris, Juho Härkönen; Daniele Vignoli and colleagues; Pew, NCFMR, ISTAT, Eurostat, IFS.
- Perspectives
- Three levels kept apart (marriage, union stability, childhood stability) and two processes (the stratification of family formation and the stratification of its consequences), tested on the United States, Northern Europe and Italy.
- Relationships
- Links to Demos (population), Paideia (education), Cadastre (housing), Labor (work) and Fides (religion and ritual).
