Amy Wax ends her 2010 account of the post-1960s family bifurcation with a description she declines to explain. Educated professionals keep bourgeois marriage and refuse to recommend it. They marry late, they marry each other, they divorce at rates that have fallen since 1980, they bear almost no children outside marriage, and they treat any public defense of the pattern as an embarrassment. Wax records the disjunction and moves on. A full treatment requires a theory of how status hierarchies select for normative regimes that benefit those at the top while imposing costs on those further down.
The theory requires three moves. The first establishes that a normative regime has an incidence, and that the incidence falls differently on populations that differ in self-regulatory capacity. The second establishes that the population selecting the regime is not the population bearing its costs, because norm production runs through institutions staffed from the upper reaches of the capacity distribution. The third establishes why the selected regime persists even when its costs become visible, and why exhortation fails as a remedy. Taken together the three moves convert Wax’s behavioral account into a political economy, and the political economy yields predictions her framework does not.
Start with incidence. Wax describes pre-1960s sexual mores as compressed heuristics, in the manner of Friedrich Hayek (1899-1992): simple rules encoding long-run knowledge about cooperation that individual actors cannot reconstruct on their own. Marry before children. Stay faithful. Honor the paternal obligation. Do not divorce. The rules relieve the follower of a calculation. Wax’s insight is that the calculation the rules perform is exactly the calculation her local thinkers cannot perform for themselves. A global thinker reaches the rule’s outcome by his own route. A local thinker reaches it only if the rule carries him there.
Run the arithmetic on who gains from enforcement. For the actor with a low discount rate, high conscientiousness, and the executive function to run a thirty-day payoff schedule in his head, the old regime delivers an outcome he reaches anyway and charges him for the delivery. It forbids a premarital sexual career he can conduct without damage. It forecloses an early marriage he might exit and replace with a better one at thirty-four. It taxes his freedom and returns him nothing he lacked. For the actor at the other end of the distribution, the regime delivers an outcome his own calculus cannot reach, at a price he was never in a position to collect. The rule is progressive in incidence. It transfers from the self-regulating to the poorly self-regulating, and the currency of the transfer is conduct.
Deregulation reverses the sign. Once the rules go, the high-capacity actor collects his freedom and keeps his outcome. He runs the sexual career, marries at thirty-two, and raises children in an intact home. The low-capacity actor collects the freedom and loses the outcome. This is what Wax means by the privatization of a cognitive burden, and it is why the egalitarian impulse behind sexual liberation produced an inegalitarian result. The sexual revolution was a regressive transfer, conducted in the language of emancipation, from those with weak self-command to those with strong self-command.
The second move asks who selected it. Here the standard accounts fail in opposite directions. One holds that the revolution was a mass movement, the spontaneous preference of a whole population that got what it wanted. The other, in the manner of Nancy MacLean, holds that a small group planned it. The first cannot explain the distribution of costs. The second requires evidence of coordination that does not exist.
The needed premise is small. Norms are produced and transmitted by institutions, and those institutions are staffed from a narrow slice of the population. Universities set the terms on which family structure can be discussed. Courts write the law of divorce, custody, and illegitimacy. Publishing houses, film studios, and networks distribute the scripts. Seminaries train the clergy who decide what the churches will still call a sin. Professional associations certify the therapists, social workers, and counselors who advise the population that most needs advice. Journalists select which claims about family life can be uttered. Every one of these institutions recruits from the top of the distribution of exactly the traits Wax identifies: verbal ability, future orientation, self-command, the capacity to project alternative scenarios. The median producer of normative content sits far above the median consumer on the trait that determines whether the norm is needed.
Each producer consults his own experience and reports it. He finds the old rule unnecessary in his own life, because it is. He finds it constraining, because it is. He observes that he and his friends conduct their sexual lives with discretion and land in stable marriages, and he concludes that the rule was superstition, since the outcome survives its repeal. He generalizes from a sample of one to a population that does not resemble him on the relevant trait.
Call this the selection problem in norm production. Any normative regime is chosen by a body whose members are systematically unrepresentative on the trait that determines the regime’s value. The regime that emerges will be fitted to the capacities of the producers rather than to the capacities of the population it governs. The result resembles a legislature that sets the speed limit by consulting only professional drivers. The professionals report that they can take the curve at ninety.
The new attitude should appear first among the populations closest to the norm-producing institutions and then move down. Fred Pampel finds the pattern in General Social Survey cohorts born between 1900 and 1985. Education predicted tolerance of premarital sex, the association strengthened through cohorts born in the early twentieth century, peaked around the 1937 cohort, then flattened and ran to zero or reversed among the youngest. Pampel reads it as diffusion of an innovation from high to low socioeconomic status, which is what Judith Treas had found earlier in the education gradient on nonmarital sex. The people least dependent on the old rule abandoned it first.
Three further engines lock the regime in place.
The first is a collective action problem of the kind Mancur Olson (1932-1998) described. Enforcement is a public good. The stigma that made the old rules work required individuals to bear private costs: the awkwardness of judging a colleague’s divorce, the reputational risk of saying that a child does better with a father in the house, the social demotion that attaches to anyone who moralizes about sex. The benefit of enforcement diffuses across a population that includes almost nobody the enforcer knows. The cost concentrates on him, in the room, at the time. Every individual therefore has reason to withhold enforcement while hoping others supply it, and the equilibrium level of enforcement falls below what even the enforcers privately believe correct. Deregulation requires only that nobody wants to pay for regulation at the margin.
The new regime carries a meta-norm. The old rule punished the conduct; the new rule punishes the punishment. Do not bear children outside marriage becomes do not stigmatize those who do. Cass Sunstein (b. 1954) describes norms as systems held up by shame, reputational incentive, and cascade, and the second-order norm runs on the same equipment as the first. The sanction did not disappear. It changed target. This raises the price of enforcement above what the collective action problem alone would set, because the prospective enforcer is not declining to supply a public good in neutral territory. He is committing an offense.
The second engine is positional. Fred Hirsch (1931-1978), in Social Limits to Growth, distinguished goods whose value survives universal provision from goods whose value depends on scarcity. The pattern Wax describes among elites, selective deviation followed by conformity, belongs to the second category. It marks status because it demonstrates a capacity most people lack: the ability to enjoy the freedom without succumbing to the disorder. Publish the recipe and the marker degrades. Defend the rule in public and one equalizes the cultural permission to depart from it, which raises the cost of departure for everyone whose self-command is weaker and lowers the distinction available to those whose self-command is strong. The interest here is in keeping a distinction that only works while it is uncommon.
The family literature reached the positional result by another route without supplying the reason. Andrew Cherlin (b. 1948) describes American marriage as deinstitutionalized, losing its binding force while gaining symbolic weight, and calls it a marker of prestige and personal achievement. That is a strange thing to happen if deregulation merely freed everyone from an obsolete arrangement. It follows if the old institution supplied a behavioral technology to everybody, and the new regime leaves marriage as something achieved by those who can supply the technology themselves. A universal rule became a positional accomplishment.
The third engine is tacit. Wax notes in passing that the well-educated may be dimly aware of why their relationships succeed. The recipe is largely unarticulated. A man who has never once considered leaving his wife for a more attractive prospect does not experience himself as exercising restraint. He experiences the question as not arising. Ask him what he does that his less fortunate countrymen fail to do and he will produce an answer about education, or opportunity, or luck, because the trait that separates them operates below the level at which he can inspect it. The failure to preach the rule is therefore overdetermined: partly interest, partly the collective action problem, partly the meta-norm, and partly that the preacher cannot locate the sermon.
Each of these engines predicts elite silence on its own, and no observation available here separates them. The essay claims their joint effect and cannot apportion it.
The distributional outcome has been described for twenty years. Sara McLanahan (1940-2021) named it diverging destinies: children of the most educated mothers gaining parental time and money while children of the least educated lost both. Cherlin named the institutional change. Ron Lesthaeghe named the ideational shift toward autonomy and self-actualization that demographers call the second demographic transition. None of the three answers the question posed here, which is why a normative transformation whose costs concentrate low in the hierarchy should be selected and maintained by institutions concentrated high in it.
Now the payoff. A theory earns its place by predicting things the theory it replaces cannot. This one predicts a pattern in what elites will moralize about.
The general decline of public moralism is a fiction. Moralism about conduct is robust and growing. The educated class moralizes with enthusiasm about smoking, seatbelts, car seats, vaccination, sugar, screen time, recycling, air travel, gun ownership, and the correct handling of a child’s disappointment. What it declines to moralize about is divorce, nonmarital childbearing, sexual promiscuity, and paternal desertion. The incidence theory explains the split. In the first set of domains, elite behavior conforms to the rule the elite recommends, and the rule costs the recommender nothing he wants. He does not smoke. He buckles his children. Recommending the rule is free, and it purchases the status that attaches to public spirit. In the second set, the rule the population needs would constrain prerogatives the recommender wishes to keep. He wants the premarital career, the option of exit, the freedom to arrange his intimate life without answering to anyone. He therefore converts the domain into a zone of private discretion and calls the conversion tolerance.
Any account that explains the decline of sexual moralism by a general rise of nonjudgmentalism has to explain why nonjudgmentalism stops at the nutrition label. This account does not need to. The line falls where the cost to the recommender begins.
A second prediction concerns insulated populations. If the regime is selected by norm-producing institutions rather than by material conditions, then a subpopulation that maintains its own norm-producing institutions should retain the pre-1960s family pattern at income and education levels where the surrounding population has lost it. Orthodox Jewish communities and the Amish supply the clearest cases. Pew reports 69 percent of Orthodox Jewish adults married against 49 percent of other American Jews, and 68 percent of married Orthodox respondents wed by twenty-four against 27 percent of other Jews. A population study of more than fifty thousand Amish households finds a median female marriage age of 20.9, 87.1 percent married by fifty, and 4.3 percent of first births conceived before marriage, with divorce rare.
Conservative Protestantism complicates the picture and the complication is useful. Jennifer Glass and Philip Levchak find higher divorce risk associated with conservative Protestant concentration, mediated by earlier marriage, earlier parenthood, and lower education and income. Strong norms are not sufficient. An institution can enforce early marriage successfully and thereby raise exposure to a different risk, which means the regime has to be assessed on the whole schedule of conduct it produces rather than on the strength of its enforcement.
The insulated cases carry a confound the argument should concede. These communities differ from the surrounding population in more than the autonomy of their norm production. They practice endogamy, they have different fertility, and they lose members. Amish attrition removes exactly the people least suited to the regime, which means observed marital stability is in part a survivorship result rather than a demonstration that the rules do the work. The comparison is suggestive and it is not clean.
What it does establish, against the strongest rival, is negative. If the differential in family outcomes tracked heritable capacity alone, the insulated communities should show the pattern of their nation. They do not. External rules substitute for internal capacity, and where the rules survive, the capacity differential stays submerged.
A third prediction concerns immigrant generations. Pew finds 22 percent of immigrant Latino youths married against about 10 percent of the second generation, and among Hispanic women who had recently given birth, 29 percent of immigrant mothers unmarried against 52 percent of second-generation mothers. Francine Blau and Lawrence Kahn find the unusually high marriage rates of Mexican immigrant women adjusted away almost entirely by the second generation. Under a material account the direction is backward, since income rises across generations. Under the incidence account it follows: the immigrant generation arrives carrying an intact regime from a society that never repealed it, and each subsequent generation spends more of its life inside American norm-producing institutions. Assimilation changes education, age structure, partner markets, residence, and selection along with norms, so this is a prediction the theory passes rather than a test it survives alone.
The theory also generalizes past sex. The same incidence logic governs every domain where a restriction protects poorly self-regulating actors at the expense of well-regulated ones. Gambling is the current case. In Murphy v. National Collegiate Athletic Association in 2018 the Supreme Court held that Congress could not forbid states from authorizing sports gambling, striking the federal statute on anti-commandeering grounds rather than lifting a federal ban on betting itself, and the practice has since spread through most states with the emancipatory rhetoric the sexual revolution used. The educated professional who places two hundred dollars on a playoff game twice a year gains a small pleasure and loses nothing. The costs concentrate where Wax’s model says they will. Brett Hollenbeck and colleagues, working from credit records for roughly seven million consumers, find ordinary legalization associated with a small decline in credit scores and the arrival of online betting associated with a decline of about twelve points, alongside rises in bankruptcy, collections, and delinquency. A 2026 working paper estimates a 2.1 percent fall in food sufficiency among working-age adults without college degrees, equivalent to roughly 10.5 percent among active bettors. A study published this year finds online legalization raising household credit-card debt 11.9 percent overall, 22.5 percent where the household head lacks a degree against 10.2 percent where he holds one. Marijuana liberalization, consumer credit expansion, and payday lending follow the same structure. In each case the restriction is repealed by a class that can handle the liberty, in the name of a population that cannot, and the repeal is described as a removal of paternalism.
Three objections deserve answers.
The first is that the repeal was not an elite act. No-fault divorce spread through state legislatures during the 1970s, and California’s Family Law Act, which Ronald Reagan (1911-2004) signed in 1969, passed a legislature answering to a mass electorate. Mass demand for divorce liberalization existed among people who were not norm producers. Elite institutions did not manufacture the demand. They removed the countervailing pressure, supplied the vocabulary in which the demand could be voiced as a right, and declined to defend the restriction when it came under challenge. A norm regime dies when nobody with standing defends it.
The second objection is that elites also bear costs. They do. Intensive parenting is expensive and anxious, delayed fertility carries a biological penalty that falls hardest on educated women, and the sexual market in the years before marriage is not costless to the people in it. The regime is not a free ride. The claim is comparative: elites bear smaller costs and collect larger benefits than the populations below them, and a comparative claim survives the observation that the costs are nonzero.
The third objection is that religious institutions did most of the enforcing, and their retreat has causes of its own that owe nothing to status interest. Secularization has its own long history. The answer is that seminaries are elite institutions too, that mainline Protestantism liberalized its sexual teaching through educated clergy trained in universities, and that the denominations which resisted are the ones whose clergy were trained outside the national credentialing system. The pattern of which churches held the line maps onto the pattern of which churches produced their own norms.
What follows for repair. The prescription that Charles Murray (b. 1943) offered in Coming Apart, that the upper class should preach what it practices, asks individuals to bear a concentrated private cost for a diffuse public benefit, in a domain where their status interest runs the other way, where the meta-norm makes the preaching itself an offense, and where they cannot articulate the practice they are being asked to preach.
A regime shifts when the composition of the norm-producing institutions changes, or when the incidence changes in a way the producers feel. Both have partial precedents. The recent willingness of demographers and economists to state the two-parent advantage in print, after two decades of treating it as unspeakable, suggests the professional cost of saying it has fallen. Whether that reaches the transmission belt below the journals is the question, and the answer so far is that it has not.
How the theory could be wrong. It predicts that permissive change appears first among high-education groups and diffuses down, which Pampel supports and which a domain showing the reverse gradient would damage. It predicts weaker diffusion where a population retains autonomous norm-producing institutions, which the Orthodox and Amish cases support and which the conservative Protestant divorce result partly cuts against. It predicts that deregulation does more damage where private resources for self-regulation are thinner, which the gambling numbers support. It predicts continued moralizing wherever the recommended conduct costs the recommender little, which the nutrition label supports. And it predicts that when the costs of a deregulation begin landing inside elite networks, elite attitudes toward regulation change.
That last one is a standing bet rather than a retrospective fit, and sports gambling may settle it. If addiction, adolescent betting, debt, and marital damage accumulate visibly among educated professionals and their children, the incidence shifts, and the theory says the language of liberty and anti-paternalism should weaken while respectable demands for restriction return. If the elite consensus holds while the damage stays concentrated below, the theory says the deregulation persists. Either result is observable within a decade without anyone running a study.
Wax writes that the populations bearing the costs have no individual route back to the arrangements that protected them. The deeper problem is that they have no collective route either, because the collective route runs through institutions they do not staff, in a register they were not trained to speak, on a question that the people who control the microphone leave closed.
Every actor in the story behaves reasonably from where he sits. The regime that results serves the people who sit highest, and it does so through the ordinary operation of who gets to speak.
Notes
On the diffusion signature. Fred C. Pampel, “Cohort Changes in the Social Distribution of Tolerant Sexual Attitudes,” 95 Social Forces 753 (2016), covers General Social Survey cohorts born roughly 1900 to 1985 and finds the education effect on tolerance of premarital sex rising, peaking near the 1937 birth cohort, then flattening to zero or reversing among the youngest, which he reads as diffusion from high to low socioeconomic status. The earlier statement is Judith Treas, “How Cohorts, Education, and Ideology Shaped a New Sexual Revolution on American Attitudes toward Nonmarital Sex, 1972-1998,” 45 Sociological Perspectives 267 (2002). Treas established the education gradient and the cohort succession; the specific premarital-sex curve belongs to Pampel.
On the positional result. Andrew J. Cherlin, “The Deinstitutionalization of American Marriage,” 66 Journal of Marriage and Family 848 (2004), is where marriage becomes a marker of prestige and personal achievement. Sara McLanahan, “Diverging Destinies: How Children Are Faring under the Second Demographic Transition,” 41 Demography 607 (2004), supplies the distributional outcome. Ron Lesthaeghe, “The Unfolding Story of the Second Demographic Transition,” 36 Population and Development Review 211 (2010), supplies the ideational account, and Zaidi and Morgan, “The Second Demographic Transition Theory: A Review and Appraisal,” 43 Annual Review of Sociology (2017), sets out its empirical weaknesses. All three describe what happened. None asks why the institutions that produced the change sit above the populations that paid for it.
On the meta-norm. Cass R. Sunstein, “Social Norms and Social Roles,” 96 Columbia Law Review 903 (1996), treats norms as systems maintained by shame, reputational incentive, norm entrepreneurs, and cascades, which is the apparatus the second-order norm runs on. The essay’s claim is that the machinery was retained and repointed rather than dismantled.
On insulated populations. Pew Research Center, “A Portrait of American Orthodox Jews” (2015), supplies the Orthodox marriage figures. The Amish population study, covering more than fifty thousand households, is at https://pmc.ncbi.nlm.nih.gov/articles/PMC12909094/. The complication is Jennifer Glass and Philip Levchak, “Red States, Blue States, and Divorce,” 119 American Journal of Sociology 1002 (2014), which finds conservative Protestant concentration associated with higher divorce through earlier marriage, earlier parenthood, and lower education and income. Anyone pressing the insulated-population argument further should confront endogamy, differential fertility, and attrition before treating these communities as a natural experiment.
On immigrant generations. Pew Research Center, “Second-Generation Americans” (2013), and the family chapter of the 2009 Latino youth report at https://www.pewresearch.org/race-and-ethnicity/2009/12/11/viii-family-fertility-sexual-behaviors-and-attitudes/. The econometric version is Francine D. Blau and Lawrence M. Kahn, “Gender and Assimilation Among Mexican Americans,” NBER Working Paper 11512 (2005).
On gambling. Brett Hollenbeck, Poet Larsen, and Davide Proserpio, “The Financial Consequences of Legalized Sports Gambling,” working from a credit panel of roughly seven million consumers. Xiaohui Guo, Lizhong Peng, and Chad Meyerhoefer, “Wagering the Bread Money: Sports Betting Legalization and Food Sufficiency,” NBER Working Paper 35305 (2026). Mustafa Enes Caliskan, “Online Gambling Legalization and Household Credit Card Debt,” Journal of Gambling Studies (2026), which carries the education split. The opinion in Murphy, 584 U.S. 453 (2018), is at https://www.supremecourt.gov/opinions/17pdf/16-476_dbfi.pdf; the statute at issue prohibited states from authorizing sports gambling and provided civil enforcement, and the Court invalidated it under the anti-commandeering doctrine rather than lifting a federal ban on individual betting.
Further reading. On the political economy of moral regimes, Hayek on the transmission of rules nobody designed, Olson on the undersupply of goods whose benefits diffuse, and Hirsch on scarcity as the source of value, all cited above. On the American family transition itself, Murray’s Coming Apart (2012), Robert Putnam’s Our Kids (2015), and Melissa Kearney’s The Two-Parent Privilege (2023) carry the descriptive record. On what elites practice and decline to preach, W. Bradford Wilcox and the National Marriage Project reports are the closest thing to a running count. On the underlying behavioral claim, Wax’s own “Diverging Family Structure and ‘Rational’ Behavior” remains the document this essay extends, and her Race, Wrongs, and Remedies (2009) is where she sets out why she thinks the arrangements that protected the losing population cannot be rebuilt from below.
