The Abundance That Wasn’t Built To Last

I. The Conditions Were External

The postwar American economy did not grow because Americans were more virtuous, better-educated, or harder-working than their successors. It grew because every serious industrial competitor had been physically destroyed and because the international financial architecture had been deliberately designed to keep things that way long enough for the United States to consolidate its dominance. This is not a hostile reading. It is the reading the Federal Reserve’s own historians give of the period.

Begin with Bretton Woods. In July 1944, while the war was still being fought in the Pacific, delegates from forty-four allied nations met at a hotel in New Hampshire and agreed on a global monetary system in which the U.S. dollar was convertible to gold at $35 per ounce and all other currencies were pegged to the dollar. Inside the architecture was a feature most people now forget. Bretton Woods permitted capital controls. Unlike the classical gold standard, the new system “was permitted to enable governments to stimulate their economies without suffering from financial market penalties.” Capital could not flee. Money could not move at the speed of a panic. A worker who won a wage increase through a steel strike in Pittsburgh was not, in the next quarter, watching the steel mill relocate to a country where the labor was cheaper, because the financial plumbing required to do that did not exist. Not because the technology was missing — the technology was perfectly available — but because the political system had deliberately not built it.

The decision to cage the dollar was the central architectural fact of the postwar American middle class, and it is almost completely absent from the popular story. Americans tell each other that the boom of the 1950s and 1960s was the natural result of national vigor, hard work, and the entrepreneurial energy of the world’s freest economy. They do not tell each other that one of the indispensable preconditions of the boom was that capital was not free.

Then there was the bombing. The United States emerged from the Second World War controlling roughly two-thirds of the world’s gold and producing roughly half of all global industrial output. European productive capacity had been destroyed twice in a generation. Japanese industry was rubble. The Marshall Plan and the parallel Dodge Line in Japan are now remembered as humanitarian reconstruction, and they were that, but they also bought the United States twenty-five years before its competitors could reach production-quality parity. Judith Stein, in her economic history Pivotal Decade, makes the uncomfortable observation that “free trade was the snake in the postwar Garden of Eden” — that from the mid-1940s through the 1970s, American policymakers “willfully allowed allies to protect their domestic markets from American goods while opening U.S. markets to their products.” In other words, the postwar trade regime subsidized European and Japanese reconstruction out of the future earnings of American workers. For one generation, that subsidy was invisible because American workers were simultaneously capturing an unprecedented share of global manufacturing rents. When the subsidy stopped being invisible — in the 1970s, when the bombed-out competitors came back online — the rent was gone, and there was nothing in the policy architecture left to defend the wage.

And then there was the energy. The postwar growth model was an energy model. Industrialized states began “doubling their energy use approximately every dozen years” after 1945, fueled by oil priced at roughly $2 to $3 per barrel. The French called the resulting boom les Trente Glorieuses — the Thirty Glorious Years. They knew they were inside a window. Americans, on the whole, did not. They thought they were inside a country.

In October 1973, the Organization of the Petroleum Exporting Countries ended les Trente Glorieuses in a week. The embargo raised oil prices by about 400 percent in a matter of days and threw the world’s economy into a sharp recession. United States consumer-price inflation jumped from 3.4 percent in 1972 to 12.3 percent in 1974. The energy-subsidized American manufacturing model had its main fuel line cut, and there was no second fuel line. Every postwar abundance argument that ignores the oil price is arguing about something else.

So: bombed competitors, caged capital, cheap oil. Three external conditions, none of them produced by anything the typical American worker did or could do, all three of them indispensable to the boom that the typical American worker came to think of as the natural fruit of working hard at a manufacturing job. None of the three was permanent. None of the three was earned. All three started to come apart in the same decade. They came apart in the same decade because they had been linked all along.

II. The Treaty of Detroit

On May 23, 1950, the United Auto Workers and General Motors signed a five-year contract that Fortune magazine, in a moment of unguarded honesty, called “The Treaty of Detroit.” The contract was not radical. It was the opposite of radical. UAW president Walter Reuther agreed that General Motors retained “all managerial prerogatives, including production and investment decisions,” and that the union would not strike during the contract’s duration. In exchange, GM agreed to a 20 percent rise in wages over five years, a $125 monthly pension beginning at retirement, a cost-of-living adjustment, and health insurance the company would pay for half. The contract’s significance was not that anyone had won a fight. The significance was that someone had agreed to a peace.

Reuther’s strategy was called pattern bargaining. The UAW would pick one of the Big Three automakers, threaten a strike, and let the other two absorb its sales until the targeted company conceded. The terms won at one company set the pattern for the next round at the others. By the time of Reuther’s death, more than half of all collective bargaining agreements in the United States contained a cost-of-living adjustment, and more than a third provided for pensions. Pattern bargaining was extraordinarily effective and, in the longer arc, strategically fragile. It worked because the UAW was capturing rents from a manufacturing sector that had no international competition. When that condition changed, the strategy stopped working, and no new strategy ever replaced it. The same is true of the rest of the labor architecture the period built. Its power was contingent on the conditions, and when the conditions evaporated the power evaporated with them, and very few of the Americans who benefited from the architecture noticed in time to defend it.

Private-sector union density in the United States peaked at 33.5 percent in 1954. One in three private-sector workers belonged to a union. By 2024, that number had fallen to 6.0 percent — from one in three to one in sixteen. The slope of the decline is sharper than the population can hold in mind. Most American workers under forty have never worked under a collective bargaining agreement and could not name a coworker who has. The institutional muscle memory of how a workplace gets organized, how a strike gets called, how a contract gets negotiated, has been lost in two generations.

For comparison: roughly 35 percent of German workers today are still covered by a collective bargaining agreement. Austria, Belgium, Denmark, Finland, France, Italy, the Netherlands, and Spain all retain coverage rates of 80 percent or more, because their laws make sectoral agreements legally binding for all firms within the sector. Japan established a universal healthcare and pension system that mixes tax and insurance fees and remains in place. Scandinavia preserved a social-democratic welfare state that did not depend on bombed competitors at any point. The American collapse from one in three to one in sixteen is not a law of physics. It is a country’s policy choice, made by a coalition that knew exactly what it was choosing.

The wage gains that the Treaty of Detroit and its imitators captured were not the result of postwar workers being more valuable than their grandchildren. They were the result of a specific institutional arrangement that distributed manufacturing rents to labor in a moment when no alternative labor supply could discipline those wages. That arrangement is gone, the rents are gone, and the institution that captured them is six percent of what it was. Mark Blyth’s phrase comes back here, the one he uses to ask, of any economic frame, what political work the frame is doing. The frame in which postwar wages are described as the natural reward of effort is doing the political work of disqualifying any conversation about the institutional preconditions that made the reward possible. The frame insists that the wage was earned by the worker. The history insists that the wage was earned by the institution that captured the rent and distributed it. The two claims sound similar in casual conversation. They are the opposite of each other in a labor-market analysis.

III. The GI Bill, and Whose

The Servicemen’s Readjustment Act of 1944 — the original GI Bill — is the single most important education policy of the twentieth century, and it is also the single most important case study in how a “universal” program in a Jim Crow polity is not universal.

The scope first. Of sixteen million American veterans returning from the Second World War, 7.8 million used the bill’s education and training benefits. Of those, 2.2 million attended college or graduate school and 5.6 million pursued vocational training in trades like auto mechanics, electrical wiring, and construction. Between 1944 and 1952, the Veterans Administration backed nearly 2.4 million home loans. The mechanism was simple: tuition plus $50 a month for subsistence (or $75 with dependents) for up to forty-eight months, plus a federally guaranteed mortgage on terms the private market would not have offered. It was the largest direct federal subsidy of upward mobility in American history, and at industrial scale, and it worked. Suzanne Mettler’s book Soldiers to Citizens documents the long shadow it cast: civic engagement rates among GI Bill beneficiaries remained measurably higher than among non-beneficiaries decades later. The program built a generation of first-in-family college graduates whose presence reshaped the structure of American civic life.

And now the qualification. The GI Bill was written race-neutral. It was administered through the existing Veterans Administration field offices, which were locally controlled and, in much of the country, segregated. The historian Ira Katznelson has spent a career documenting this story, and the dollar-figure receipts he produces are the part of the postwar mythology most carefully kept out of the family album.

In 1947, in thirteen Mississippi cities studied by Katznelson, only two of more than 3,200 VA-guaranteed home loans went to Black borrowers. The denominator is not a typo. Two. Out of more than three thousand two hundred. In the New York and northern New Jersey suburbs — the same suburbs the GI Bill was busy financing into existence as a form of social architecture — 67,000 GI Bill mortgages were insured, but fewer than 100 of them went to non-whites. Historically Black colleges and universities, which under the prevailing legal regime were the only institutions Black veterans in the South were permitted to attend, were chronically underfunded and could not absorb the volume of incoming Black veteran applications. Many of those veterans simply could not use the benefit they had earned by serving in the same army as the white veterans who used theirs.

The standard academic framing is that the GI Bill “widened an already huge racial gap.” That is correct as far as it goes. A more precise way to put it is that the GI Bill was the down payment on white suburban middle-class formation, and Black veterans were the mortgagors whose down payment was deposited into someone else’s account. The white suburb of 1965, with its single-family homes and its property-tax-funded schools and its inheritable home equity passed forward through the generations, was financed in part by a federal program that excluded Black veterans from the same benefits, in the same year, on the basis of the same service. Any honest accounting of the postwar abundance has to mark that exclusion. The mythology that “everyone” had a chance is not historically accurate. It is historically accurate that some specific people had a chance, and that other specific people were systematically excluded from the program designed to give them one, and that the gap between those two groups was opened by federal policy at the moment of greatest opportunity.

This is not a digression from the abundance story. It is the abundance story. The white middle class of 1965 and the Black middle class that did not form on the same schedule are two ledgers in a single accounting, and the abundance of the first ledger is partly explained by the deliberate underwriting of the second. The series’ second article walked through the legal and judicial sequence by which segregation and disinvestment were maintained in the schools. This article’s job is to add the postwar pivot point: the moment when the federal government, through the most generous education and housing program it had ever passed, sorted Americans by race at the precise moment of greatest mobility.

IV. The Tuition Table That Wins the Argument

The National Center for Education Statistics publishes a historical table called Digest 330.10. It is a single-page document, freely available on the NCES website, listing the average tuition and required fees for in-state undergraduates at American public four-year institutions for every academic year going back to 1963–64. There is no chart in American higher-education policy that does more damage in a smaller number of cells.

Two readings. First, the nominal price went up by a factor of about 32 over the period covered by the table. Second — and this is the reading that matters — the inflation-adjusted price went up by a factor of about four and a half. The growth in real cost is not an accounting illusion produced by inflation. It is a real, multi-fold expansion in what attending a public four-year college actually costs an American household relative to the goods and services that household needs to live.

Set the price against income. In 1964–65, the median family income in the United States was about $6,569. Tuition was about 3.9 percent of one year of household income. By 2020, the average cost of tuition and fees at a public four-year institution represented over 35 percent of median household income, up from approximately 18 percent as recently as 1999. In the least affordable states the number is grotesque: Pennsylvania at 72.5 percent, Rhode Island at 71.2 percent, New York at 68.3 percent. Roughly nine times the relative burden in fewer than two generations.

The collapse is not because state governments stopped funding higher education in absolute terms. They did not. Per-student state appropriations adjusted for inflation were roughly $7,447 in 1980 and $11,683 in 2024. The story is in the composition. In 1980, state appropriations accounted for about 79 percent of total revenue at public universities. By 2019, that share had declined to about 55 percent. By 2016, tuition revenue exceeded state and local funding for higher education in half the states. The fall is in the share of cost the state was willing to absorb on behalf of the student. The state did not slash its higher-education budget. The state failed to expand its higher-education budget in proportion to enrollment, cost inflation, and the marginal cost of the new programs the political system had asked the universities to add. The marginal cost was shifted to the student. The shift was performed quietly, over forty years, in budget meeting after budget meeting, and the result is that the public university stopped being a public good and started being a publicly licensed private good, charging private prices, financed by private debt, defended by the rhetoric of public access.

The rhetoric is the part that does the political work. The university is still called “public.” The website of the same flagship state institution that in 1965 charged a few hundred dollars and was understood to belong to the state’s residents now charges several thousand dollars and is described in the same brochure language. The continuity of the language masks the discontinuity of the financial arrangement. A student who arrives at the university having been told that her parents and grandparents went there for nothing, or for almost nothing, is told now that the same institution will cost her or her family the equivalent of a down payment on a house, and is invited to understand this transition as the natural growth of an institution that has only become more excellent over time. The language is doing what the budget did. Both pieces are pretending that nothing has changed about what the institution is for.

V. The Work-Through-College Math

Now combine the two columns of the receipt at the top of this article. In 1965, the federal minimum wage was $1.25 an hour. A full-time forty-hour-a-week summer job, worked across fourteen weeks, cleared $700 before any payroll deductions. Tuition at the average public four-year college was $256. The summer job covered tuition 2.7 times over, and there was money left over for books, transportation, and a contribution to room and board.

The historian Randal Olson has done the careful version of this calculation across the full span of NCES data. In 1965, his finding is that a student working at the federal minimum wage during the school year would only have to work about 23 hours a week — roughly 3.3 hours a day — to cover the full cost of attendance at a public four-year institution. That number is not aspirational. It is the number of hours of part-time work a 1965 student needed to cover what she owed the school, while still having time for classes, study, and a social life. The arithmetic was loose enough to permit a college experience.

The 2022–2023 version of the same calculation. A student working at the federal minimum wage now needs 26 hours a week, year-round, just to cover tuition and fees at the same public four-year institution — before paying for housing, food, textbooks, or transportation. Twenty-six hours per week, year-round, comes to roughly 1,510 hours of work per year. The same arithmetic that in 1965 produced 205 hours produces 1,510 hours in 2022. The change is not a small number. It is an order of magnitude.

Two notes about that number. The first is that it covers tuition and required fees only. It covers no part of the cost of having a body that needs to sleep somewhere, eat something, and read books. The second is that it assumes the student works straight through the academic year at the federal minimum wage, which means she is not also a full-time student in any meaningful sense. The 1965 student could work during the summer and study during the year. The 2022 student must work during the year and study around the work. She is not a student with a job. She is a worker with a class schedule.

None of this is a moral indictment of the 2022 student. She is doing exactly what the 1965 student did — selling labor at the prevailing minimum wage to pay for the prevailing tuition at the prevailing kind of public university. The difference between them is not effort. It is the ratio between minimum wage and tuition. The ratio collapsed because the minimum wage stagnated and the tuition expanded, and neither of those things was the student’s fault, and neither of those things was the result of the student’s deficient work ethic. They were the results of policy. They were the policy results of a series of decisions about how much of the cost of higher education the public would absorb on behalf of its residents and how much the residents would be asked to absorb on their own. Those decisions had names. They had budgets. They had legislative records. They are public, and they are recoverable, and they are not a story about anyone’s laziness.

The “work your way through college” story that Boomers tell their grandchildren is therefore not a moral story. It is a ledger entry describing a specific ratio of minimum wage to tuition that existed for roughly thirty years and does not exist now. The ratio inverted, and with it the feasibility of the entire debt-free graduation narrative. When the grandchildren listen to the story and try to repeat it — the summer job at the gas station, the part-time job in the cafeteria, the paid-off tuition at the end of August — the arithmetic stops them somewhere in the second week. They are not failing the grandparents’ test. They are taking a different test, in a different room, on a different curve, with a different proctor. The same name on the cover sheet is the only thing the two tests share.

VI. The Iceberg Underneath the Wage

The wage was never the whole compensation. Postwar compensation is systematically under-measured by wage data alone, because so much of what a 1965 factory worker earned came in the form of benefits the employer supplied as a condition of employment and the union enforced through the contract. This is not a footnote. It is the thing.

Among medium and large private firms in 1979, 87 percent of full-time workers participated in a retirement plan, and most of those plans were defined-benefit pensions. A defined-benefit pension is a contractual promise from the employer to the worker that, after a certain number of years of service and at a certain age, the employer will pay the worker a fixed monthly amount for the remainder of the worker’s life. The promise is from the employer to the worker. The investment risk and the longevity risk sit on the employer’s balance sheet. The worker is told what she will receive when she retires, and she receives it.

From 1980 through 2008, the share of private wage-and-salary workers participating in defined-benefit plans fell from 38 percent to 20 percent. By the mid-2020s, only 11 percent of private-industry workers were in defined-benefit plans at all. The defined-benefit pension, which was the bedrock of postwar retirement security for the largest cohort in American history, has been functionally discontinued for the workforce that is supposed to inherit the promise of the same labor architecture.

What replaced it was the 401(k). The provision was added to the Internal Revenue Code in the Revenue Act of 1978, in a section drafters did not even particularly notice at the time, intended originally as a supplemental tax-deferred savings vehicle for already-pensioned executives. Through a regulatory interpretation by Ted Benna in 1980 and a quiet expansion through the 1980s, the 401(k) became the dominant vehicle by which American private-sector workers were expected to fund their retirement. The expansion was not voted on. There was no national debate about whether to abolish the defined-benefit pension. There was a series of accounting decisions, regulatory rulings, and corporate cost-cutting moves through which one form of retirement security was substituted for another, and the substitution was complete by the time most workers noticed it.

The substitution is not cosmetic. A defined-benefit pension is a promise from the employer to the worker. A 401(k) is a tax-advantaged brokerage account that the worker is expected to manage on her own, bearing all market and longevity risk. The risk that the stock market will be in a trough on the day the worker retires sits on the worker. The risk that she will live longer than she planned sits on the worker. The risk that her employer will go bankrupt and stop matching her contributions sits on the worker. The risk that her own discipline as a saver will fail her sits on the worker. The risk transfer from employer to worker is total, and it happened during the lifetime of the Boomer generation — the same generation that received the defined-benefit plans their children and grandchildren will never see. By 2018, more than half of private-industry workers had access only to defined-contribution retirement plans. There was no defined-benefit option to opt into. The promise had been retired before the worker had a chance to accept it.

Lawrence Lessig’s reading of architecture is the relevant analytical move here. The four modalities Lessig names — law, norms, markets, architecture — describe how behavior is regulated. ERISA, the 1974 statute meant to protect pension promises, ended up — through a series of subsequent regulatory rulings — providing the legal scaffolding inside which the 401(k) substitution was performed. The architecture changed. The rhetoric did not. Workers are still told to “save for retirement” as if it were a personal virtue, when what has actually happened is that an institutional risk previously borne by the employer has been quietly relocated to the worker’s household balance sheet. The personal-responsibility framing is doing the political work that Mark Blyth’s framework predicts: it makes a structural transfer look like a moral test. It is much easier to ask whether an individual worker “saved enough” than it is to ask why the employer no longer bears any portion of the risk it bore one generation ago.

Health insurance is the parallel story. Employer-provided health insurance became the dominant American form of healthcare coverage almost by accident — a wartime workaround for World War II wage controls, institutionalized by tax policy, and locked into the labor contract by the Treaty of Detroit generation. As long as unions were powerful enough to extract it and employers were profitable enough (and trapped in domestic markets) to provide it, the system functioned. The share of private-sector workers with employer-provided health insurance has been falling since the early 1980s. The Affordable Care Act in 2010 was the federal government belatedly acknowledging that the employer-based system was no longer covering the people who needed it.

The compounding effect is the part that wage statistics miss. A 1965 factory worker’s nominal hourly wage of $4.00 represented an hourly wage plus a defined-benefit pension plus employer-paid health insurance plus a cost-of-living adjustment plus seniority-based job security. A 2025 warehouse worker earning $20 an hour in nominal dollars has none of these. Any comparison that treats the two compensation packages as roughly equivalent because of CPI adjustment is comparing non-comparable bundles. The benefits iceberg that Boomers received and Millennials do not is worth, conservatively, 30 to 40 percent of the wage on top. It is the dog that doesn’t bark in the wage-stagnation statistics, and it is the reason that the wage-stagnation statistics under-report the actual gap in compensation between the two generations.

VII. The Chart That Breaks the Mythology

If you have to look at one chart in American labor economics, look at this one. The Economic Policy Institute publishes it under the title “The Productivity-Pay Gap.” It is reproduced in nearly every major labor economics textbook, and it has been replicated in independent calculations using BLS data, Census data, and Federal Reserve data. The shape it shows is the most important shape in the recent American economic record.

From 1948 to 1973, the hourly compensation of a typical American worker grew almost exactly in lockstep with labor productivity. The two lines are, for all practical purposes, the same line. As the economy became more productive — as American workers and the capital they used produced more output per hour — the compensation of the typical worker rose with it. Productivity went up. Pay went up. The two were tracking. The contract, in the political and moral sense, was holding.

Then the lines diverge. Between 1973 and 2014, productivity grew 72.2 percent, or about 1.33 percent annually. The compensation of the typical worker grew 9.2 percent, or about 0.22 percent annually. The most recent EPI update finds that productivity has grown roughly 3.5 times as much as typical worker pay since 1979. That ratio — 3.5 to 1 — is not a noisy estimate. It is robust to how you measure productivity, how you measure pay, which deflators you use, and which industries you include. It is the central finding of forty years of careful empirical work, and it is the receipt that ends the meritocracy argument.

Where did the productivity gains go? Into capital, not labor. Thomas Piketty’s central finding in Capital in the Twenty-First Century — that the top 1 percent and especially the top 0.1 percent have captured a rising share of national income since roughly 1980 — is the other side of the same coin. The EPI decomposition attributes the productivity-pay gap to rising inequality, with the top 10 percent and especially the top 1 percent gaining a much larger share of all compensation, and labor’s share of income eroding since 2000. The productivity that the typical American worker helped to generate did not disappear. It went somewhere. It went to people who were not the typical American worker.

The decoupling is not a law of physics. It is a policy outcome. The key policy decisions can be listed in chronological order, and each one has a specific institutional author who can be named.

1971–1973. Nixon closes the gold window in August 1971, ending the dollar’s convertibility to gold. Bretton Woods collapses. Capital controls begin to erode. Global finance opens. The financial plumbing the postwar settlement depended on is dismantled in stages.

1973. The OPEC oil embargo ends cheap energy. The energy-subsidized growth model loses its main subsidy. Les Trente Glorieuses are over.

1979. Paul Volcker is appointed Chair of the Federal Reserve and announces a new operating procedure aimed at breaking the inflation cycle. By June 1981, the federal funds rate has been raised to 19.1 percent. By late 1982, unemployment has reached 10.8 percent. The recession of 1981–1982 is the deepest postwar downturn to that point. Manufacturing employment in “Rust Belt” states — Michigan, Ohio, Pennsylvania — is hit hardest. About 90 percent of the job losses in that recession occur in mining, construction, and manufacturing. Many of those jobs never come back.

1981. Reagan fires 11,345 striking air traffic controllers in the PATCO strike, sending an unmistakable signal to the entire private sector that the federal government will no longer enforce union protections in any politically inconvenient labor dispute. PATCO is not a one-day event. It is the structural permission slip for the next forty years of corporate union-busting.

1981–1986. The Reagan tax cuts shift the tax burden down the income distribution and reduce the top marginal rate from 70 percent to 28 percent. The federal government’s capacity to redistribute through the tax code is structurally narrowed.

1980s. The deregulation of finance, airlines, and trucking expands. The financialization of the U.S. economy — the share of corporate profits coming from financial-sector activity rather than from production — begins the long climb that Judith Stein documents in Pivotal Decade.

The Volcker shock in particular is the decisive event. It was framed as an inflation-fighting measure, and as an inflation-fighting measure it worked — inflation dropped from double digits to near-trivial within a few years. But the distributional effect of the shock was a sledgehammer to industrial labor. Manufacturing jobs lost in the 1981–82 recession in the Midwest never returned during the recovery. Jefferson Cowie, in his history Stayin’ Alive: The 1970s and the Last Days of the Working Class, names the dynamic precisely: “Labor insurgencies got benefits but no control over the system, leaving the organization of production untouched until it was radically reshaped in the Volcker Shock at the end of the decade.” The 1970s is the decade in which the postwar labor settlement was dismantled. By 1985 the machine that had produced the middle class was already being demolished. By 2000 the demolition was complete.

This is the material base of what Sandel diagnoses as the moral corrosion of meritocracy. When the mythology says “work hard and you will be rewarded” but the data say productivity grew 3.5 times faster than pay since 1979, the mythology does not merely become inaccurate. It becomes an instrument of humiliation for the people who believed it and did the work anyway. Sandel’s “tyranny of merit” is the psychological superstructure of the post-1973 wage stagnation. The tyranny is what it feels like, from the inside, to have done what you were told and to have not received what was promised, and to be told the absence of reward is itself a measure of your character.

VIII. The Fading American Dream

Raj Chetty and his collaborators published the definitive measurement of the collapse in absolute income mobility in Science in 2017. The headline finding is single sentence in length and I will reproduce it here without compression.

Rates of absolute mobility have fallen from approximately 90 percent for children born in 1940 to 50 percent for children born in the 1980s.

Translate that into a human frame. A child born in 1940 had roughly a nine-in-ten chance of out-earning her parents at age 30. A child born in 1984 — a Millennial, in other words, the cohort the present article is in some respects writing for — had roughly a fifty-fifty chance. The mythology that “every generation does better than the last” became statistically false during the generation-one transition from Boomers to Generation X, and the trend line continued downward through every cohort that followed. The decline is across the entire income distribution, with the largest declines for families in the middle of it. This is not a story about the rich getting richer at the expense of the very poor. It is a story about the middle losing the upward escalator the previous generation rode.

Robert Putnam reaches the same conclusion through ethnography rather than statistics in his book Our Kids: The American Dream in Crisis. His method is to walk the streets of the town he grew up in — Port Clinton, Ohio — and to talk to the families who lived there in 1959, when his own high school class graduated, and to the families who live there now. In the 1959 Port Clinton he describes, the vast majority of students went on to lives better than their parents’. Social class mattered relatively little for getting ahead. The high school football team had the children of the bank president and the children of the steel-mill foreman on the same field, and they all expected to do something with their lives that involved a step up from where their parents were. In the Port Clinton of the present, the opportunity gap has widened dramatically — partly because affluent kids now enjoy more advantages than affluent kids did then, but mostly because poor kids now are in much worse shape than their counterparts were then. Putnam’s book is not nostalgic. He is careful to note that 1959 Port Clinton was racially segregated, that the women of the same generation were largely excluded from the upward mobility their brothers experienced, and that the steel mill at the center of the town’s economic story closed in the 1970s. Port Clinton is not paradise. Port Clinton is a town that briefly approximated the mythology and now does not, and the gap between what it was and what it is can be measured in classmates’ lives.

Tyler Cowen, working from a different ideological starting point, reaches a compatible conclusion. His phrase for the period after 1973 is the “Great Stagnation,” and his observation that median family income more than doubled between 1947 and 1973 but increased by less than one quarter between 1973 and 2004 is doing the same numerical work that Chetty’s mobility data is doing. Cowen’s frame is the “complacent class” — the class that won the meritocratic game and has organized its life around protecting its winnings. The series will return to that frame in the next article. For now, the relevant point is that Cowen’s numbers and Chetty’s numbers and Putnam’s ethnography all converge on the same finding: a turning point around 1973 after which the postwar pattern of broad-based intergenerational improvement breaks, and the breaking is not a temporary downturn but a structural shift in the relationship between effort and reward.

The cruelty of the situation is intergenerational. The Boomer cohort that benefited from the postwar conditions did not author them. The Boomers were the children of the Greatest Generation, born into the architecture the Greatest Generation built and the postwar conditions made possible. They inherited the conditions. They told their own children that the conditions were a law of nature. The children, taking the parents at their word, tried to repeat the work and were met with the receipts above. They are not lazier than their parents. They have a different ratio.

The Habermas frame, lightly applied, is useful here. The lifeworld of working-class community deliberation about labor, dignity, and intergenerational obligation has been progressively colonized by financial-planning instruments that treat retirement as a personal optimization problem, college as an individual investment decision, and household financial security as a function of personal spreadsheet hygiene. Conversations that used to happen at union halls and at kitchen tables — about whether the contract was fair, about whether the work was being valued, about what was owed to the children — have been displaced by conversations about how much to put in the 401(k) and which student loan repayment plan to choose. The displacement is not absolute, and Habermas’s argument is not that the financial-planning instruments are evil. The argument is that they have crowded out a different kind of discourse, and that the kind of discourse they crowded out was the only discourse in which the structural questions could be asked. Inside the financial-planning frame, every problem becomes an individual optimization problem. The structural conditions that made the optimization easy or hard to begin with become invisible. They become invisible because the frame is incapable of representing them, and the frame is the only one most households now have available.

IX. The Demographic Trap

The Baby Boom itself — 76 million births between 1946 and 1964 — was a one-time demographic event, and the postwar abundance was partly its product and partly its beneficiary. When a large young cohort enters the workforce, the worker-to-beneficiary ratio for retirement programs improves mechanically, which reduces the tax burden on the working population and permits higher take-home pay. When the same cohort retires, the mechanics reverse.

The Social Security Administration’s own projections are explicit on this point: “The worker-to-beneficiary ratio is fairly stable in years the boomers are in the workforce (1980–2005) but is substantially lower when the boomers are in their retirement years (2020–2040).” The Congressional Research Service confirms the same story. The cost of Social Security will soon begin to increase faster than the program’s income, because of the aging of the boom cohort, expected continuing low fertility, and increasing life expectancy. None of these trends is a moral failing. They are all demographic mechanics.

The cruelty of the mechanics is intergenerational. The same cohort that benefited from a demographic dividend during its working years — paying relatively low Social Security taxes because the retiree population was small — is now collecting benefits from a demographic burden the working population has to finance. Millennials and Generation Z are being asked to pay a higher share of their wages into programs from which they will not receive the same real benefit. The intergenerational politics of this are radioactive, and they are downstream of demographic math, not of any moral failing on either side. The Boomers are not freeloaders. The Millennials are not whiners. They are inhabiting the opposite ends of a demographic curve neither of them designed and the timing of which neither of them chose. The math is the math. The conversation about it should sound that way.

X. Could the Conditions Be Rebuilt?

The strongest counter-argument to the “conditions cannot be replicated” thesis runs as follows, and it deserves a fair hearing because parts of it are correct and the article should engage it rather than dismiss it.

The argument is this. The postwar abundance was not primarily the result of geopolitical luck. It was primarily the result of deliberate policy choices — a coalition of decisions about taxation, public investment, financial regulation, organized labor, and social insurance that produced an embedded liberal settlement of the kind Karl Polanyi described. The top marginal income tax rate was above 90 percent in the early 1950s. Public investment in higher education was generous. The GI Bill, the Federal Housing Administration, and the Wagner Act were each major federal interventions in markets that otherwise would not have produced the postwar middle class. Antitrust enforcement was vigorous. Monetary policy aimed at full employment. Corporate management accepted constraints on its discretion in exchange for labor peace. None of these conditions were imposed by physics or geography. They were chosen. And the reason they eroded after 1973 is that they were unchosen — defunded, deregulated, and disenforced — not because the world changed, but because a coalition of interests decided to dismantle them. Adam Tooze frames this as the “polycrisis” of an embedded settlement that is now visibly in crisis but whose dismantlement was political rather than natural. Gary Gerstle’s The Rise and Fall of the Neoliberal Order is even more explicit: the New Deal order “lasted from the 1930s to the early 1970s” and was “undergirded by the simple but powerful idea that a strong interventionist state was necessary to regulate capitalism.” Gerstle treats the neoliberal order that succeeded it as a contingent formation, not a natural one, and he notes that it is now visibly in crisis — which opens the door, in principle, to a successor order that could in theory choose differently.

If the conditions were policy choices, the steelman argues, they can in principle be chosen again. A sufficiently ambitious reform program — Medicare for All, free or near-free public college, sectoral collective bargaining on the German or Belgian or Nordic model, progressive wealth taxation, full-employment monetary policy, antitrust revival, serious industrial policy — could in principle reconstruct something like the postwar settlement without depending on bombed competitors or cheap oil. The international comparative evidence is the strongest piece of ammunition for this view. Germany, Japan, and Scandinavia also ran postwar abundance economies, and they have preserved more of the settlement than the United States has. Germany retains roughly 35 percent collective bargaining coverage. The Scandinavian countries have preserved social-democratic welfare states with single-digit child poverty rates. Japan established universal healthcare that mixes tax and insurance fees and remains in place. None of these countries has bombed-flat competitors or two-dollar oil either. The difference between their outcomes and the American outcome is policy and institutional design, not raw materials.

The steelman’s strongest numerical point is this: if you control for national policy choices, the variation in 2025 outcomes across the high-income world is larger than the variation explained by geopolitical and energy conditions. Sweden in 2025 is closer to Sweden in 1965 than the United States in 2025 is to the United States in 1965, and the reason is institutional design, not natural endowment.

This is a serious argument and the article wants to take it seriously. It is correct on the architectural question. The settlement was chosen. A new settlement could in principle be chosen. The article concedes that. The series concedes that. There is nothing about the post-1973 American outcome that is metaphysically inevitable. The country that ran progressive marginal tax rates above 90 percent and a 33.5 percent unionized private-sector workforce in the 1950s could in principle choose to do that or something like it again. The political coalition required to make those choices is hard to assemble in the contemporary American polity, but the obstacle is political, not physical.

Where the article does not concede is on the magnitude question. The 1965 ratios — tuition at 3.9 percent of household income, a summer job covering tuition 2.7 times over, a 33.5 percent private-sector union density, an 87 percent defined-benefit pension coverage rate — were partly the product of the external conditions the steelman cannot reproduce. The bombed competitors were a one-time event of the Second World War. The cheap oil was a function of a global energy market that no longer exists at $2 to $3 per barrel and never will again. The caged dollar was a feature of a financial architecture that has been deliberately taken apart and could not be reassembled in a single legislative session even by the most ambitious reform government, because the global capital flows it once contained have grown to a scale at which capital controls would have very different consequences than they did in 1948. A reformed United States, even one that adopted every plank of the most ambitious progressive program currently on offer in serious policy circles, would not recover the 1965 ratios. It might approach Swedish ratios. It might do considerably better than its current trajectory. It might restore something that looks like a middle class to households that no longer feel they belong to one. None of those would be small accomplishments. All of them would be improvements over the present, and all of them are worth fighting for. But none of them would be the thing the postwar mythology claims as its baseline.

The mythology is a two-part claim. First, that the postwar abundance was the natural product of American virtue. Second, that if you work hard the way the postwar generation did, you will experience the same abundance. The steelman undermines the first claim, and rightly so. It does not rescue the second. Even under the most generous reading of what is achievable through political reform, the material conditions that produced the 1965 ratio cannot be replicated identically. They can only be approximated by a very different set of policies in a very different economic environment, and the approximation is being refused by the very people who benefited from the original. The gap between “rebuildable” and “rebuildable to 1965 levels” is itself a measure of what the mythology obscures, and it is a much larger gap than the casual telling of the story permits.

The honest position, then, is something like this. The postwar settlement was contingent and chosen. It can be chosen again, and a country that chooses it would be considerably better off than the present American arrangement permits. But the magnitude of what was briefly achieved in the postwar United States depended on conditions that were not chosen and are not available to be chosen again. The gap between “a much better country than this one” and “the country your grandparents lived in” is real, and refusing to mark that gap is the way the mythology keeps doing the political work it was built to do.

XI. The Conditions, the Mythology, and the Bill

Tyler Cowen’s frame for the period after 1980 is the “complacent class.” The series’ sixth article will spend more time inside that frame. For the closing of this article, the relevant move is the one Cowen makes about “matching technology.” Cowen argues that the post-1980 era has produced a series of new technologies — the Common Application for college admissions, Zillow for residential search, the various dating apps, the streaming-music recommendation engines — whose effect has been to make sorting more efficient. Each technology promises choice and delivers efficient segregation. Each one helps the user find people, places, and things very similar to herself, with the result that the cross-class encounters that used to happen by accident in American life now happen mostly by mistake. The matching technology is the dynamic complement to the post-1973 wage stagnation. The wage gap created the material divide. The matching technology made the divide easier to inhabit. The class that benefited from the conditions has organized its life around protecting the position the conditions purchased and around minimizing accidental contact with the people for whom the conditions did not work.

This is the “complacent class” in Cowen’s phrase. It is not the top one percent. It is the upper 30 to 40 percent of earners — the credentialed professional class whose children attend the better-funded school districts, whose homes are zoned to exclude denser housing, whose families are connected to the network of internships and references that move the next generation through the credential pipeline. They are the beneficiaries of the postwar conditions transmitted down two or three generations, and they have, in aggregate, organized their lives around making sure the same conditions persist for their children and not for everyone else’s. They are not villains. Most of them are decent people doing what every parent in every culture has always done, which is try to give their children the best start they can. The political problem is what happens at the level of the country when the best start that several million decent parents secure for their children is structurally available only to them, and is then defended by the rhetoric of effort.

The rhetoric of effort is the load-bearing piece of the mythology. It is the part that allows the people who inherited the conditions to tell themselves that the conditions were earned. It is the part that allows the same people to tell their grandchildren that the grandchildren’s difficulty in repeating the inheritance is a measure of the grandchildren’s character. The arithmetic at the top of this article is the answer to that rhetoric. Two hundred and five hours, then. Fifteen hundred and ten now. The gap is not character. The gap is the conditions, and the conditions are gone.

Sandel’s contribution to this article is the moral diagnosis. The post-1973 wage stagnation is what produces the material humiliation. The mythology that says “you should have made it if you tried” is what turns the material humiliation into a moral verdict on the person who failed to make it. Together, the wage stagnation and the mythology produce the political condition Sandel calls the tyranny of merit. It is a tyranny because it is inescapable from inside its frame. The person inside the frame cannot tell the difference between a structural condition and a personal failing. She has been taught a vocabulary in which there are no structural conditions, only personal failings, and the vocabulary is doing its political work whether or not she can name what it is doing.

The work of this article has been to make the structural conditions visible enough that the vocabulary loses some of its grip. The arithmetic at the top of the article is the most important sentence the article contains, and it is doing the most important work the article can do, which is to put a number on the gap between two generations and to make the number speak for itself. Two hundred and five. Fifteen hundred and ten. The same country, the same kind of school, the same federal minimum wage in real dollars per hour of human time. The difference between the two numbers is everything the rest of this article has been describing, and the difference is not a difference of character. It is a difference of conditions.

The conditions were the bombed competitors and the caged dollar and the cheap oil. The conditions were the Treaty of Detroit and the 33.5 percent union density and the GI Bill that the federal government wrote race-neutral and administered through Jim Crow field offices. The conditions were the public university funded at 79 percent by the state and the defined-benefit pension that the employer guaranteed and the wage that grew with productivity because the institutional architecture forced the productivity gains to be shared. The conditions were a particular accident of geopolitics and a particular set of institutional choices, and they were both less natural and less permanent than the people who benefited from them ever bothered to notice.

When the conditions evaporated, the pipeline carried only debt. The university kept charging tuition. The employer kept the productivity gains. The pension was discontinued. The wage stayed flat. The mythology kept telling the children of the people who had received the conditions that the conditions were a law of American nature. The children inherited the story and the debt, but not the conditions.

This is the part of the series’ argument that is closest to the surface of contemporary American life. Almost everyone who reads this article is, in some way, inside the receipt at the top of it. They are either the people who lived through the conditions and remember the ratio that no longer exists. Or they are the people who inherited the mythology of the ratio and are trying to repeat it against an arithmetic that will not let them. Or they are the people watching the second group struggle and wondering whether the second group is failing or whether something else is happening that the available vocabulary does not let them name.

The available vocabulary does not let them name it. That is what the mythology is for. The job of this article is to put a number on the thing the vocabulary obscures, and to leave the number sitting at the top of the page where the conversation about it has to start.

Two hundred and five.

Fifteen hundred and ten.

Read them as a ratio. The ratio is the entire story.

The next article in this series moves from the public university to the private elite university, from the partial postwar opening to the four-century reproduction engine that briefly admitted a few of the postwar veterans and then went back to doing what it had always done. The receipts are different. The arithmetic is the same kind.