The Family Curve
Marriage, households, and the shape of everyday life are changing faster than at any point since the industrial revolution. The data on what is actually happening, before the politics of what should.
(28 for women, 30 for men - up from 21 and 23 in 1960)
(up from about 5% in 1960)
(up from about 6% in 1980)
A note on framing. Family is a topic where what people feel about the data often runs ahead of the data itself. The numbers below describe what has changed - they do not say whether the change is good or bad. Reasonable people of every political persuasion can look at the same statistics and reach different conclusions about what they mean. The page tries to walk through the most important measured shifts, what is contested, and what the practical consequences are for individual readers and for whole societies.
What has actually changed
The shape of daily family life has shifted dramatically in every developed country and increasingly in middle-income countries too. The shifts are remarkably consistent across very different cultures, which suggests they are driven by deep structural changes rather than by any single political or religious decision. Several of the largest changes:
Marriage is later, and increasingly optional. The median age at first marriage has risen by roughly seven to nine years in most developed countries since 1960. American women now marry at a median age of 28 (up from 21); Japanese women at 30 (up from 24); Italian women at 33 (up from 24). The share of adults who never marry at all has also risen sharply - in the US, about 28% of 40-year-olds today have never been married, up from roughly 6% in 1980.
Childbearing has separated from marriage. About 40% of US births today are to unmarried mothers. In France the figure is 60%; in Iceland and Norway it is over 70%. In 1960 it was below 10% in almost every Western country. The change is not primarily about teenage motherhood (which has actually fallen) - it is about adult women in long-term cohabiting relationships having children without formal marriage.
Cohabitation has risen sharply. Living together without marriage has gone from being a minority practice to a near-universal stage of relationships in most developed countries. About two-thirds of US first marriages today are preceded by cohabitation. In France and the Nordic countries, formal marriage has become increasingly optional - many couples raise children together for decades without ever marrying.
Divorce rates rose, then fell. US divorce rates roughly doubled from 1960 to 1980, peaking around 1980. They have been falling slowly since then. Today the US divorce rate is back near the 1970 level. The story of "divorce keeps rising" is a 1980s story that newer cohorts are not living - though the marriage rate has fallen so much that the comparison is not entirely apples-to-apples.
Single-parent households have stabilised at a high level. About 23% of US children today live with a single parent, mostly mothers. That share roughly tripled between 1960 and 1990, then plateaued. Most other developed countries have followed similar patterns.
Family size has fallen at every income level. Connected to the demographics page on this site, the average woman in every developed country now has fewer than two children, and most are well below 1.7. Whether this reflects deliberate planning or constraints (housing costs, childcare costs, career considerations) is hotly debated; both stories have evidence behind them.
Why these shifts happened
No single cause explains the family curve. The changes are best understood as the cumulative result of several deep shifts that all happened roughly together starting in the 1960s and 1970s.
Women's education and work. The single biggest variable across all developed countries. As women got more education and entered the workforce in larger numbers, the financial and life-path reasons to marry early or at all changed dramatically. Women whose grandmothers had to marry at 22 to be financially secure could now choose differently and increasingly did.
Reliable contraception. The pill was approved in the US in 1960 and became widely available through the 1960s and 1970s. For the first time in human history, sex and reproduction could be cleanly separated. The downstream effects on dating, marriage, and childbearing took decades to fully play out and are still playing out.
The decline of the marriage premium. When marriage was the only socially acceptable path for adult women and the only stable way to raise children, the cost of staying single or divorcing was extraordinary. As social norms shifted and welfare systems expanded, the practical cost of being unmarried fell sharply. The change is not primarily ideological - it is structural, and it removed a powerful stabilising force.
Economic stress at the bottom. Marriage rates have fallen most sharply for non-college-educated Americans. The reasons are debated but several converge: men without college degrees have seen their real wages stagnate or fall over forty years, making them less attractive marriage partners by traditional standards; women across the income distribution have gained education and earning power; and the kind of stable industrial jobs that supported the working-class American family in 1960 have largely disappeared. Marriage has become, increasingly, a pattern of educated and higher-income households.
Cultural and religious shifts. Most developed societies have become more secular, more individualistic, and more tolerant of family forms that would have been stigmatised in 1960. The mechanisms here are real but harder to measure. Religious participation correlates with more stable marriage across developed countries, though disentangling causation from selection (religious people differ from non-religious people on other dimensions that also predict stability) is genuinely hard. Whether secular substitutes can do the same job is one of the open questions of the next generation.
How countries actually compare
The shifts have happened in every developed country but the pattern, pace, and resulting family forms differ. The numbers below describe rough national averages and are approximate.
The takeaway: the direction of change is broadly similar across developed and middle-income countries, but the speed and the resulting forms vary widely. France and the Nordic countries have settled into a stable cohabitation-with-children pattern; the US has fragmented along class lines; East Asia has trended toward marriage-and-children-becoming-rare-and-late; Mediterranean Europe has changed slowly; much of Africa, the Middle East, and South Asia is on different trajectories with different starting points.
The paths from here
Family forms tend to change slowly, in part because they are reproduced through how each generation is raised. The next twenty years are likely to look more like extensions of current trends than sharp breaks, but several different shapes are possible.
Continued slow drift
Marriage rates keep falling in most developed countries; cohabitation continues to expand; the age at first marriage and first birth keeps rising; the class gap in family stability stays wide. By 2040 marriage is a minority adult experience in most cohorts, though stable long-term partnership remains the norm.
Will it happen? This is the base case. The forces driving the trends - women's education and economic opportunity, cultural shift toward individual choice, falling religious participation, and economic stress at the bottom - are all still operating. None look likely to reverse soon.
Stabilisation at current levels
The most extreme decline countries (South Korea, Italy, Spain) hit a floor as the social and economic costs of very-low marriage and birth rates become severe enough to provoke serious response. Marriage rates stabilise rather than continuing to fall. Other countries follow similar patterns at higher levels.
Will it happen? Possible. The signal from South Korea (singletons becoming a near-majority of households, severe demographic decline) is now visible enough to be a political issue. Whether it stops the trend or just slows the trend further is unclear. Most family-system researchers think a partial stabilisation is more likely than a real reversal.
Class-based polarisation deepens
The American pattern continues: highly-educated and higher-income families increasingly look stable and traditional, while less-educated and lower-income families increasingly raise children outside of marriage with high partnership instability. The two patterns become permanently divergent rather than the lower group catching up to the upper.
Will it happen? The US has been on this path for forty years and there are no obvious signs of convergence. Several other developed countries (UK, Australia, parts of Eastern Europe) show signs of similar polarisation. The class-stratified family is a structural feature, not a temporary one.
New forms of partnership emerge
Living-apart-together arrangements (couples who maintain separate households), legal civil partnerships short of marriage, multi-generational households, and intentional co-parenting arrangements among friends all become more common. The "marriage versus cohabitation" frame becomes less helpful as the actual variety of stable adult arrangements expands.
Will it happen? Already starting. France's PACS (civil partnership), Japan's growing single-person household share, and the rise of multi-generational households in expensive American metro areas all point in this direction. Whether these become the new norm or stay marginal depends on legal recognition and social acceptance.
Pronatalist policy bends the curve
Governments increasingly worried about demographics offer larger and more sustained packages of support for marriage, childbearing, and family life - direct cash benefits, parental leave, child allowances, housing subsidies, and tax preferences. France, Hungary, Japan, and South Korea have all spent meaningfully on this. Effects so far have been modest but real (about 0.1 to 0.2 children per woman).
Will it happen? Partly already happening. Most developed countries have expanded family support over the last twenty years. The size of the increase needed to fully reverse the trend is much larger than what has been tried so far. A combination of stronger support and broader cultural change might bend the curve more than support alone.
Religious revival changes the family
Religious communities (in the US, increasingly in parts of Eastern Europe, in much of Latin America and Africa) maintain traditional family structures with notably higher fertility and stability. As secular fertility falls, the share of children raised in religious households grows. By the late 21st century the developed-world religious-secular composition could shift visibly.
Will it happen? Already happening at the margin. Eric Kaufmann and other demographers have argued for years that the children of more religious parents will reshape the population over generations. The effect is real but slow; it shifts the average over decades, not over election cycles.
Migration shapes the family-form picture
Migrant communities in developed countries often have higher fertility and more traditional family structures than the surrounding population. Over generations, the descendants tend to converge toward the local pattern, but the convergence takes time. The composition effect on family-form averages depends substantially on migration policy and integration patterns.
Will it happen? Already happening visibly in major immigrant-receiving countries. The full size of the effect depends on how much migration each country accepts and how migrant communities settle.
The realistic forecast is, again, a mix. Continued slow drift is the base case for most developed countries. The class gap in stable family structure is unlikely to close without serious economic intervention at the bottom. Some stabilisation in the most extreme decline countries is plausible. The new forms of partnership are quietly already filling some of the space that traditional marriage used to occupy. None of this is likely to look much like 1960 again - the underlying changes in women's education, contraception, and economic structure have not reversed, and the family forms produced by those underlying changes are different from the ones they replaced.
Where serious analysts disagree
Family is one of the topics where analysts agree on more of the data than the political conversation suggests, and disagree more on what the data means. Each reading below is held by named scholars whose work is worth reading directly.
Family structure matters more than the tolerance framing admits
Children raised in stable two-parent households have better measurable outcomes - educational, economic, behavioural - than children in single-parent households, even controlling for income. This is not a moral claim; it is what the data shows across multiple countries and decades. Pretending otherwise to avoid stigmatising single parents has costs of its own, particularly for the children whose outcomes are at stake.
Held by: Brad Wilcox (University of Virginia), Melissa Kearney (Maryland), and a substantial fraction of family-research sociologists. Their data on outcomes is robust; the policy implications are contested. Different researchers in this group reach different conclusions about what to do about it.
The decline-of-the-family framing is overstated
What is happening is not the decline of the family - it is the diversification of family forms. Stable cohabiting partnerships, blended families, multi-generational households, and adult-with-grandparent arrangements all play the role that marriage used to play, often with comparable stability. The "decline" framing measures only the marriage-shaped subset of what families actually are.
Held by: Andrew Cherlin (Johns Hopkins) and the broader sociology-of-the-family community. Their data on the variety and stability of non-marriage arrangements supports them; the open question is whether the average outcomes (rather than the headline labels) actually look comparable.
The class gap is the deeper story
Both the "decline" and the "diversification" framings miss the most important pattern: the family experience of college-educated, higher-income Americans diverged from the family experience of less-educated, lower-income Americans starting around 1980, and the gap has widened ever since. The two groups now live increasingly different lives, with different family stability, different parenting, and different outcomes for the next generation.
Held by: Robert Putnam (Harvard, "Our Kids"), Sara McLanahan (Princeton), and a substantial fraction of inequality researchers. The class-and-family lens has become much more central in serious research over the last fifteen years and is more useful than older framings for understanding what is actually happening.
Women's preferences explain more than the structural framings allow
Surveys of women's stated preferences about marriage, childbearing, and family size have changed dramatically over generations. Many of the family-curve trends reflect women genuinely choosing different lives than their grandmothers chose - not the simple result of structural constraints. Reducing the trends to economics, policy, or culture misses the reality that the preferences themselves have shifted.
Held by: Alice Evans, several British and European demographers, and a tradition of preference-respecting feminism. The implication is that policy aimed at "fixing" the family curve has limited room to operate without overriding what individual women actually want.
Family stability is a public-health-scale issue
The mental-health, educational, and economic consequences of family instability for children and adults are large enough that they should be treated as a major public-health concern, not as a private cultural matter. Underinvesting in the supports that promote stable family life - paid parental leave, accessible childcare, housing affordability, marriage-friendly tax policy - is a long-term cost the system is paying without acknowledging.
Held by: Kathryn Edin (Princeton), Bradford Wilcox, and a coalition of researchers from across the political spectrum. The framing is unusual because it cuts across left and right - both sides have a version of this argument with different policy prescriptions.
None of these readings is fully right or wrong. What can be said from the available evidence: family forms have genuinely shifted in ways that reflect both expanded individual choice and structural pressures; the consequences for children's outcomes are real and measurable; the consequences are distributed unequally across class and race; and the right policy response is contested among careful researchers in ways that do not map cleanly onto left-versus-right politics.
What this means for you
Family choices touch every reader directly in ways most other Now topics do not. Some practical observations that are useful regardless of one's political or religious views:
If you are choosing a partner
The data strongly suggests that the quality of a long-term partnership matters more for life satisfaction than almost any other variable - more than career success, more than wealth, more than residential location. Choosing a partner well is one of the highest-leverage decisions an adult makes. The qualities that predict stable partnerships (kindness, respect, shared values, reasonable conflict-resolution skills) are themselves more predictive than money, looks, or any of the variables that dating culture tends to overweight.
If you are deciding whether to have children
The decision is more personal than almost any other on this site, and it is yours alone. What is worth knowing: parenting is harder than most people expect, more rewarding than most people expect, and almost universally one of the things adults look back on as among the most meaningful parts of their lives. The economic and career costs are real and measurable. The emotional and developmental returns are real and harder to measure. The decision should be made on what you and your partner actually want, not on what you feel pressured to want by demographic worry, pronatalist policy, or career anxiety in the other direction.
If you are raising children
Stability of the home matters substantially - more than household structure type, more than household income above a basic threshold, more than the specific choices about education or activities. Children do well when adults around them are reliable, attentive, and emotionally regulated. They do less well when those adults are unstable, distracted, or in conflict. The single most important parenting input is reducing chronic stress in the household, which compounds over years in measurable ways.
If you are single
Long-term singlehood is now common enough that the social infrastructure for living alone successfully is much better than it used to be - financially, socially, and culturally. The mental-health consequences of long-term loneliness are real, however, and worth taking seriously. Friendship networks, community involvement, and physical health habits are particularly important for long-term singles. The "find a partner" advice that older generations defaulted to is not the only option, but the underlying need for sustained meaningful relationships is universal whether they are romantic or not.
If you are thinking politically
Family policy is one of the few areas where serious people across the political spectrum can agree on a meaningful subset of policies (paid parental leave, expanded child tax credits, child-care support) even when they disagree about deeper questions of marriage, gender roles, or religious values. The pragmatic case for supporting families regardless of structure - because the children's outcomes matter to everyone - is one of the more under-utilised pieces of common ground in contemporary politics. Voting on the merits of specific policies rather than on slogans about family decline or family freedom tends to be more useful than the political conversation suggests.
The mechanics behind this
The family curve sits on top of three deeper mechanisms covered elsewhere on this site. If the analysis above depends on ideas you want to understand first, these fundamentals make the conversation more legible:
