The marriage rate is one of the cleanest signals demography produces, and across the developed world it is pointing in one direction: down. In the United States, the crude marriage rate has fallen from about 9.8 per thousand population in 1990 to about 6.0 in 2022. In Japan, the rate has fallen from 6.4 to 4.1 over roughly the same period. In South Korea, from 9.3 in 1990 to 3.7 in 2022 — a brutal collapse that has accompanied a fertility rate now at 0.72, the lowest ever recorded for a sovereign state. Across the European Union, the average marriage rate has fallen from 6.3 in 1990 to roughly 4.3 today, with Italy, Spain, and Portugal at the low end and France pioneering the substitution of cohabitation and civil partnership for legal marriage.
The 2nd Law — Think — asks for clear sight of what these numbers actually mean. They do not mean that the desire to pair has vanished; partnership in various forms persists, and in some countries (France notably) the legal-marriage decline has been partly offset by civil-union arrangements. They do mean that the institution of marriage as the default adult arrangement, which it was across most of the 20th century in all four regions, is no longer the default. Marriage has become one option among several — postponable, conditional, and increasingly elective rather than expected.
The drivers are partially shared and partially region-specific. Shared across all four regions: rising female educational attainment and labor force participation, which raises the opportunity cost of early marriage; the contraceptive revolution, which decouples sex from marriage and marriage from childbearing; the rise of cohabitation as a socially acceptable alternative; declining religious affiliation, which removes a normative pressure toward marriage; and the smartphone-and-app transformation of the dating market discussed elsewhere in this volume.
Region-specific drivers add texture. In the US, Andrew Cherlin and Philip Cohen have documented a sharp class divergence: college-educated Americans still marry at high rates, while marriage has retreated dramatically among non-college Americans, producing a stratified pattern where marriage has become both rarer and more class-coded. Phil Cohen's 2019 paper "The Coming Divorce Decline" complicated the narrative by noting that the people who do marry now are more selected, better matched, and divorcing less — but the price is that fewer marriages happen at all. The US story is one of marriage retreating to a privileged subset.
In Japan, Mary Brinton's work documents the structural mismatch between an inflexible labor market that demands male long-hour commitment and a domestic culture that demands female full-time homemaking. Young Japanese women, increasingly educated and employed, find the available marriage bargain unattractive: marry and exit the workforce, or remain single. Many choose the latter. Japan's shōshika — the low-birthrate problem — is built on a marriage-rate collapse that the state has been unable to reverse despite three decades of policy attempts.
South Korea presents the same dynamic in concentrated form. Sang-Hun Choe's reporting in the New York Times over years has chronicled the multiple feedback loops: extreme housing costs, the cost of cram-school education, gender-role rigidity, a winner-take-most labor market, and a feminist-vs-anti-feminist political polarization that has made cross-gender solidarity politically fraught. The result is a marriage rate that has more than halved and a fertility rate that demographers describe as "below extinction" — meaning that if current rates persist, the Korean population will halve within two generations.
European patterns are more varied. France and the Nordic countries have substituted cohabitation and civil partnership for marriage to a substantial degree, with fertility rates that have held up better than the marriage rates alone would suggest. Italy, Spain, and Portugal have not made this substitution — partnership outside marriage remains less common — and they show the lowest combined marriage-and-fertility rates in Europe. The mediating variable seems to be whether the society has built functional non-marital partnership institutions or has left marriage as the only legitimate adult-pairing form, in which case the marriage decline is also a partnership decline.
At the collective scale, the consequences extend across generations. Fewer marriages mean fewer households formed earlier, fewer children, fewer multigenerational care networks, and a thinner kinship infrastructure to support aging populations. The economic implications — for pensions, for housing demand, for labor supply, for healthcare — are large and approaching. The cultural implications — for what adulthood means, for what daily life looks like, for what kind of stories the society tells about itself — are larger still and harder to measure.
What does the 2nd Law ask of this analysis? It asks for resistance to two equally easy errors. The first is the conservative error: that the marriage decline is a moral collapse to be reversed by exhortation. The data does not support exhortation as a lever; in every country that has tried it, the rate has continued to fall. The second is the progressive error: that the decline is pure liberation with no costs, and that the consequences for fertility, loneliness, and demographic structure are either fabricated or irrelevant. The data does not support that view either. The honest position is to hold both: real gains in autonomy and real losses in social cohesion, with the trade-off unevenly distributed across class, gender, and region, and the policy levers that would shift the balance largely un-pulled.