Meet three Americans with student loan debt.
The first is a newly minted Harvard MBA. He owes $300,000. He just signed an offer for $350,000 a year. He will pay off his loans in roughly the time it takes most people to pick a paint color.
The second got a four-year degree in a “studies” field from a school whose career-services office turned out to be a fiction. She owes $20,000, works as a retail clerk, and has no clear path to earning the kind of income that would retire the debt. She defaulted years ago and has been hounded by collections ever since.
The third is the unemployable graduate of a for-profit college that was shut down by the Justice Department for fraud. She owes $3,000. The training she paid for prepared her for a career that didn’t exist the way the brochure described it.
When you read that 42.8 million Americans hold roughly $1.7 trillion in federal student debt, you are reading a sentence that mashes those three people together. You are also reading a sentence that can be used to argue for almost any policy you like—from forgiving the MBA’s $300,000 because debt forgiveness is “popular,” to expanding federal lending so the next round of students can pay even higher tuition at schools designed to capture the loans.
This kind of mashing is the basic move of the contemporary affordability discourse. Catastrophic situations affecting small numbers of people are blended with mild discomforts experienced by very large numbers, and the resulting mixture is called a crisis. The numbers grow large because the categories are elastic. Then the crisis is used to justify policy interventions that, more often than not, make both the small severe problem and the large mild one worse.
There are real affordability problems in the United States. Two of them, by our count. The rest is noise, and the noise is doing a lot of political work—for both parties, in different ways. None of it produces good policy and most of it makes the genuine problems harder to solve.
Most affordability writing sorts Americans by income decile or sorts the world by service category—the costs of housing, healthcare, or childcare. We are doing something different. We are sorting Americans by the kind of trouble they are actually in. There are two groups in real trouble: a destitute tail that is small and brutally squeezed, and a much larger squeezed-talent class that is doing everything right but still cannot replicate its parents’ material life. Everyone else is fine, or fine enough, or in the kinds of trouble we are not obligated to call a crisis. The two groups need opposite remedies, and the “single homogeneous affordability crisis” framing makes that impossible to see.
The ‘49 percent.’
In March 2026, the Urban Institute released an updated version of its True Cost of Economic Security measure. The headline finding is that 49 percent of people in American families lack the resources required for economic security. The Urban Institute itself is careful in using the term economic security, not survival, and notes that more than 40 percent of those below the threshold sit between 75 and 100 percent of it—“largely getting by, and meeting most regular expenses, but not primed to thrive.” But the coalition that commissioned the report—the National True Cost of Living Coalition—and the press that covered it weren’t careful. Fortune in March: “[H]alf of Americans can’t afford to dine out or vacation in a cost of living crisis.” Governing: “Nearly Half of U.S. Households Are Not Economically Secure.” The Community Service Society of New York, one of the report’s commissioners, in its own release: “More Than Half of All People in U.S. Lack Economic Security.” The careful researcher’s number becomes the activist’s slogan becomes the political claim.
If that were true, you would expect to see it. You would expect mass starvation, or at least mass migration out of the country. You would expect the streets to be lined with the dispossessed. What you actually see is a country whose biggest dietary problem is obesity and whose biggest housing problem is that nobody is allowed to build any. So what is the 49 percent number actually measuring?
The threshold is defined as the level of resources required to fully participate in today’s economy and society without cutting back on anything, while also saving for emergencies and the future.
To fail this test, you do not need to be hungry. You do not need to be cold. You do not need to be one car repair from homelessness. You need only to be cutting back on something. Or not saving as much as some Urban Institute economist thinks you should.
By that standard, the Harvard MBA from our cold open fails the test. He’s spending all $350,000, plus borrowing more, because he just bought a co-op in Manhattan and his fiancée wants a destination wedding. He is cutting back—on dessert at Per Se. He is not saving for the future. By the True Cost of Economic Security definition, he is part of the 49 percent of Americans who cannot afford to live.
So is a retired couple in Florida who are deliberately drawing down their savings, which is what savings are for. So is a 23-year-old barista who’s spending every nickel on rent and concert tickets and having the time of her life. So is the unemployable woman with the worthless degree from the fraudulent for-profit college who is genuinely choosing between insulin and electricity in rural Mississippi.
One number. Four completely different stories. The number is engineered to include all of them, because including all of them is what produces the 49 percent. Even that number, taken at face value, doesn’t tell you the share of Americans in actual distress. It tells you the share who feel a financial squeeze of any kind, of any severity, for any reason.
The two real problems.
There are two serious affordability problems in America. They look nothing like each other and need opposite solutions.
The first is the destitute tail. A real but small share of Americans are in genuinely miserable financial situations. They have more bills than they can pay. They are one missed paycheck from eviction. They frequently have literally zero money. The unemployable woman with the worthless degree from the fraudulent for-profit college is in this category. So is the 58-year-old who got laid off from a manufacturing job, exhausted his savings, can’t get hired anywhere, and watches his wife work double shifts at Walmart.
These people need money. The institutions that make their lives worse—the for-profits that produce unemployable graduates, and the medical billing systems designed to confuse people into paying twice—need to be regulated or eliminated. Both of those statements are true, and neither is in serious political dispute.
The second problem is the squeezed-talent class, and it’s harder to explain because the people involved look fine on paper. Picture a 32-year-old physician married to a 32-year-old software engineer. Combined household income, $400,000. They cannot buy a house in San Francisco or Boston or New York within a sane commute of their jobs. They cannot afford to have three kids, pay for childcare, and put them in decent schools. They are doing every single thing the meritocratic American dream told them to do, and the dream is not being delivered. Their parents, at the same age, with worse credentials and lower real incomes, owned a house and had three kids on one salary. Something is broken here, and it isn’t their fault, and it isn’t fixed by transfers. Giving this couple a $5,000 childcare credit doesn’t move the needle on $4 million houses—and worse, by raising effective demand for childcare without doing anything about the supply, the credit makes childcare more expensive for the people behind them in line. The right tends to dismiss this couple as coastal-elite complainers. The left tends to dismiss them because they’re already in the top 5 percent of incomes. Both are wrong. This is a talent-allocation problem of the first order, and a country pays a real price when its most productive young people can’t form families or live near their work.
These two problems require completely different policy responses. The first example is a redistribution problem, requiring cash transfers, safety nets, regulatory crackdowns on the worst predators. The second is a supply problem and requires zoning reform, occupational licensing reform, and breaking up the entrenched cartels in housing, healthcare, education, and childcare. The packages don’t overlap.
This is where the affordability discourse goes wrong, and where both political coalitions go wrong with it. Democrats propose policies notionally aimed at the destitute tail—childcare subsidies, student loan forgiveness, ACA expansion—that mostly pass through to the squeezed-talent class while raising the underlying prices for everyone. Student loan forgiveness is regressive by construction: The people with the largest balances are the ones who went to graduate school and earn the highest lifetime incomes. Childcare subsidies disproportionately help dual-earner professional couples, and where the subsidies are biggest, the prices have risen the fastest. Republicans propose tax cuts and deregulation that mostly help asset owners and high earners, plus, more recently, tariffs and industrial policy that hurt the destitute tail by raising prices. In short, both parties subsidize the upper middle class while assuring everyone the help is going to the poor.
The grab bag.
Now let’s look at what people actually tell pollsters bothers them. The answers fall into three tiers, and only the third is “noise” in any pure sense. The top tier is food and gas. These dominate every recent open-ended poll by a wide margin. CNN’s May 2026 survey found the share of Americans spontaneously naming gas prices as their top economic problem rose from 5 percent to 23 percent in a single year, with food costs cited almost as often. The University of Michigan consumer sentiment index hit an all-time low in May, with the survey director attributing the decline explicitly to the oil-supply shock from the Iran war. Data from the Bureau of Labor Statistics has energy up roughly 18 percent year-over-year and gasoline up 28 percent. These are commodity stories, not a structural cost-disease story. They are real and they hurt, especially the destitute tail. But for anyone not desperately poor, what they inflict is temporary budget pain and pessimism, not the structural experience of being priced out of a life. We will not pretend they don’t move the political needle.
The second tier is the structural items—housing, healthcare, childcare, and education. In open-ended polling, they show up at lower frequencies than food and gas: Gallup’s April 2026 survey asks respondents for their most important financial problem; housing came in at 13 percent and healthcare at 8 percent. This is affordability, but specific affordability for specific, relatively small groups.
The third tier is the long tail of grievances that fills out the rest. These do not lead any reputable poll, but they show up in focus groups, in op-eds, and crucially in the survey responses that drive the 49 percent number. Picking through them is illuminating, because it is here that “affordability” visibly stops meaning what it sounds like it means.
Streaming subscriptions. Yes, really. The same households that paid for cable, long-distance phone calls, and a video rental membership in 1995 now complain that they have too many $5 streaming bills. This is a discretionary substitution. It is also a quality-of-life improvement their grandparents could not have imagined.
Restaurant prices and tipping. Real, annoying, and in some cities outrageous. Americans spent about 43 cents of their food dollar in restaurants in 1990, in 2025 it had risen to 56 cents. The complaint about restaurant inflation is, in significant part, a complaint about a voluntary behavior change.
Travel youth sports. “My kid’s club soccer is $4,000 a year.” Yes, and it’s optional, and it didn’t exist in your childhood, and your kid is not going to play in the World Cup.
Pet healthcare. Genuinely up a lot, and people truly suffer over it. Also, a chosen luxury that 1970 households mostly couldn’t access at all. The MRI for the dog was simply not on the menu back then. Are people worse off because this option now exists but is expensive?
None of this belongs in a serious policy conversation. All of it shows up in the survey responses that drive the headline affordability numbers. When 49 percent of Americans say they can’t afford to live securely, some meaningful share of them are telling the pollster they had to pick between Disney+ and Netflix.
We don’t say this to mock. (Well, only somewhat to mock, and in any case the mocking is fond.) We say it because lumping in the 23-year-old who’s mad about streaming bundling with the 58-year-old former machinist who can’t afford insulin is a category error. They are not having the same problem. The first one needs to grow up. The second one needs help. It is possible to be deeply sympathetic to the second while being a little snide about the first; in fact, taking affordability seriously requires it.
A working taxonomy.
If we’re going to talk about affordability seriously, we need to sort the actual problems by what kind of problem each one is. There are roughly four categories.
The first is cheap to fix. Predatory subprime products—the for-profit colleges that produce the unemployable graduates, the title lenders, the local governments using fines as revenue sources. Surprise out-of-network billing. Some categories of occupational licensing. Most of these are fixed not by spending more money but by removing things government already does badly, or by making certain business models illegal. These are layups. They have not been hit because the small set of beneficiaries from the status quo is concentrated and motivated, while the larger set of victims is diffuse.
The second is worth real money. Pediatric healthcare for poor children. An expanded earned income tax credit. Targeted housing vouchers and subsidized childcare for actually working low-income parents—in places where the supply will respond. These expenditures buy genuine outcomes—provided, and this is the operative caveat, the supply side is allowed to respond. Allsubsidies do when supply is artificially restricted is bid up the price.
The third is “people should handle it themselves.” Lifestyle creep. Social-comparison goods. Optional enrichment activities for children. The choice to live in the most expensive ZIP codes in the country. The choice to pursue a degree with a clearly negative net present value when an honest cost-benefit calculation was available. These are not policy problems. They are adulthood. Calling them “the affordability crisis” is one of the more insidious forms of cultural infantilization going around, and we are not in fact required to agree with the framing.
The fourth is genuinely hard structural problems with no cheap fix. Housing in productive coastal metros. Eldercare. The cost-disease sectors generally—healthcare, higher education, anything where Baumol’s curse (upward salary pressure in fields without productivity gains) meets a regulatory cartel. These will not be fixed by any single intervention because the political coalition that benefits from the status quo is too large and too local. Housing in San Francisco will not become affordable until incumbent homeowners accept that the common good requires slower increases in house prices, which is to say, probably never. They got rich by closing the door behind them, and they would like that door to stay closed.
The two real affordability problems map to two different boxes. The destitute tail needs the cheap fixes plus targeted real-money transfers. The squeezed-talent class needs the structural fixes, which are not cheap politically even when they are nearly free fiscally. Almost everything else either fixes itself or shouldn’t be fixed.
The part nobody mentions.
Here is the move that the affordability discourse pulls off most quietly. The implicit assumption in essentially every affordability story is that we’re at year zero. That the problem has been ignored. That nobody has tried to do anything about it. That If Congress Would Only Act, things would get better.
Congress has been acting. Continuously. For decades. On every single one of these problems.
Take student loans, where we started. The federal government has been “addressing” student-loan affordability since the Higher Education Act of 1965. Pell Grants in 1972. Expanded loan eligibility in 1978 and 1992. Income-based repayment in 1994 and again in 2009. Public Service Loan Forgiveness in 2007. The federal takeover of the entire loan program in 2010. The SAVE plan in 2023. Broad-based forgiveness attempts from 2022 through 2024. The cumulative federal effort over those six decades runs into the trillions in subsidies, forgiveness, and forgone interest.
And what happened? Tuition rose at roughly twice the rate of general inflation for the entire period during which the federal government was “making college more affordable.” The standard explanation is that subsidizing demand for an inelastic good causes the price to rise to capture the subsidy. Schools raised tuition because students could borrow more. Students borrowed more because schools raised tuition. The aid rose to chase the price. The price rose to absorb the aid.
If you want the mechanism caught on hidden camera, here it is. In 2025 Congress set a new $100,000 lifetime cap on federal borrowing for graduate students, taking effect July 1, 2026. Within weeks, UC Irvine’s Paul Merage School of Business cut its Flex MBA tuition by $30,000—landing at exactly $99,000—and announced the new price with marketing copy reading, without apparent irony, “University of California MBA is priced within reach of government loan limits — making a world-class degree not just aspirational, but truly attainable.” The school did not need econometric identification. The school had a price-setting algorithm and a federal loan cap, and it set the price to the cap. Even the administration that wrote the cap conceded the mechanism: The stated rationale was that “unrestricted borrowing led to steep increases in graduate school tuition.” This is the Bennett hypothesis as confession, in real time.
This is not a secret. It has been documented by the New York Fed, by Beth Akers, by basically anyone who has looked at the data without an ax to grind. But the affordability discourse cannot metabolize the finding, because the finding implies that the standard policy menu—more aid, more forgiveness—is the cause of the problem, not the solution.
Healthcare is the same story at a larger scale. The ACA. Medicare Part D. Medicaid expansion. CHIP. COBRA. HSAs. The employer-provided health insurance tax exclusion, which by dollar value is the largest healthcare subsidy in the federal code and has been quietly inflating both demand and prices since 1943. The No Surprises Act. Hospital price transparency. The Inflation Reduction Act drug-pricing provisions. The federal government has done an enormous amount on healthcare affordability over the last 25 years. Some of it worked. Medicaid expansion clearly reduced uninsurance. The ACA exchanges stabilized what had been a death-spiraling individual market. Some of it failed. The lowest-cost Silver plan on the ACA marketplace has risen 77 percent since 2017, which the Urban Institute tracker leads with as evidence of the affordability crisis. That number is, in significant part, the consequence of choices made in 2010 about how the individual market should be structured. The affordability crisis in ACA premiums is the consequence of an affordability policy.
In 2021, the American Rescue Plan dramatically expanded the premium tax credits as pandemic relief; the IRAin 2022 extended the expansion through 2025. Marketplace enrollment more than doubled, from 11.4 million to 24.3 million, with 92 percent of enrollees receiving the enhanced credits. Insurers priced into a market organized around the new subsidies. Consumers chose plans they could not have chosen without them. When the enhanced credits expired at the end of 2025, the average annual marketplace premium payment roughly doubled, from about $888 to about $1,900, according to KFF. A 60-year-old couple at 402 percent of the federal poverty level—the cliff the expansion had smoothed—could see their annual premium go from 8.5 percent of income to roughly a quarter of it. The subsidy was expanded. The market reorganized around it. The subsidy was withdrawn. The “affordability crisis” in 2026 ACA premiums is, in significant part, the affordability policy of 2021 finishing its arc.
Housing is the most awkward case for both parties. Federal housing policy since the 1930s—the GI Bill mortgages, FHA, Fannie and Freddie, the mortgage interest deduction, the low-income housing tax credit, Section 8 vouchers, the Community Reinvestment Act—has been an enormous, sustained, bipartisan effort to make housing more affordable. The cumulative subsidy probably runs into the tens of trillions in present value over the postwar era. And the result is that housing in productive metros costs more relative to income than at any point in living memory.
The reason is not a mystery. Essentially all of these programs subsidized the demand for housing while local zoning rules restricted the supply. Subsidies that meet a fixed supply capitalize into the price of land. Edward Glaeser and Joseph Gyourko have spent 20 years documenting this. The YIMBY movement is built on this insight. And the affordability discourse, again, cannot quite bring itself to say it out loud, because saying it out loud requires confronting that the chief obstacle to housing affordability in America is the median voter in the most desirable places.
Childcare is the newest and least-subsidized of the major categories, but the early evidence is moving in the same direction. The Child Care and Development Block Grant was created in 1990 and has been expanded multiple times. The American Recovery and Reinvestment Act of 2009 and the American Rescue Plan Act of 2021 injected tens of billions more. States with the most generous childcare subsidies—New York is the current poster child—have seen prices rise fastest. And childcare, like housing, has a heavily restricted supply: State-level rules on staff-to-child ratios, group sizes, and worker credentialing meaningfully limit how many centers can operate. Multiple studies find that tightening the infant ratio by a single child reduces the number of centers in the average market by roughly 10 percent, with the closures concentrated in lower-income areas. Childcare in the most-regulated states is more than twice as expensive as in the least-regulated. So when subsidies expand the demand, the supply cannot respond.
Now look at the pattern. The categories where we have spent the most public money trying to improve affordability—higher education, healthcare, housing, and, increasingly, childcare—are precisely the categories where affordability has gotten worse, sometimes catastrophically so. The categories where we have done very little—food, clothing, electronics, telecommunications, most consumer goods—have gotten dramatically more affordable in real terms over exactly the same period. A 1985 long-distance phone call from the United States to Europe could cost more in nominal dollars than a month of unlimited cell service in 2025 that includes a supercomputer in your pocket.
If you want to view this entire pattern in a single image, look at the chart Mark Perry of AEI has been updating for years, the one labeled, only somewhat in jest, the “Chart of the Century.” It plots U.S. price changes from 1998 forward across 14 categories. The categories that have risen most steeply, in some cases more than tripling, are hospital services, college tuition, college textbooks, medical care, and childcare—the heavily subsidized, supply-restricted ones. The categories whose prices have fallen, in some cases by more than 90 percent, are TVs, toys, software, and cellphone service—the ones with no significant subsidy, robust supply response, and exposure to international competition. New cars and clothing, contrary to the conventional grumble, have been roughly flat in nominal terms for a quarter-century while wages have risen 80 percent or more—meaning they have become substantially cheaper in real terms. The pattern is so clean it looks like the figure was drawn to make the point.
We are not claiming that correlation proves causation. But this is well beyond mere correlation. The data show this happening in every category that fits that combination, and the opposite happening in every category that doesn’t. We are claiming that anyone who wants to argue that we just need to spend more on the affordability crisis bears a much heavier burden of proof than is typically asked of them. The track record of the standard menu is bad. In some categories it is catastrophic. The default policy reflex—more subsidies, more forgiveness, more aid—is the same reflex that produced the underlying inflation in the first place.
Why the fog is useful.
If the lumping-in is so destructive, and the policy history is so unflattering, why does the affordability discourse persist in exactly its current form? Because the fog is useful. To both parties. In different ways.
Democrats point to the 49 percent number and propose childcare subsidies, student loan forgiveness, and ACA expansion. The poor get rhetorical attention, the upper-middle class gets the dollars, and the prices keep rising. It is a remarkable feat of policy: Each round of intervention generates the political case for the next round of intervention, while the underlying conditions get worse for everyone except the producers of the subsidized service.
Republicans point to the same number and propose tax cuts and deregulation. These mostly help asset-owners and the squeezed-talent class. They do very little for the destitute tail: The capital-gains rate is not a binding constraint on a single mother in Mississippi, and the estate-tax exemption is not what’s keeping the unemployable for-profit graduate from getting a job. And this is what Republicans used to do. To this they have lately added massive tariffs and selective industrial policy. Tariffs are an explicit tax on the consumer goods the destitute tail buys; the worse the household’s situation, the larger the share of its budget that gets taxed. To call this an affordability policy requires an Orwellian flexibility of language.
Why is a flat-screen TV affordable and a college education not? … The answer is not a coincidence, and it is not because flat-screen TVs are easier to make than college degrees. It is because Congress has spent 60 years trying to make college affordable and has spent zero years trying to make TVs affordable.
The poor don’t have lobbyists, but they do have activists, and the activists are in many cases more important than lobbyists. The middle class has union leadership and institutional advocates who are also more important than lobbyists. The fog of affordability discourse is convenient for all of these actors, because it lets them claim they’re addressing a single homogeneous crisis while pursuing policies aimed at their constituencies.
The fog is also a useful way to avoid confronting the policy history. If you accept that the 49 percent number describes a single homogeneous problem, you don’t have to ask why the categories where we’ve spent the most have gotten worse. You can just keep proposing more of the same. If you separated the problems and looked at the track record honestly, you would have to ask much harder questions—and the answers might require betraying parts of your coalition. Better to keep the fog.
What honest stats would look like.
If you wanted to describe American affordability honestly, you would do a few specific things differently from the Urban Institute tracker.
You would use separate denominators for separate problems. Childcare prices matter to the roughly 15 percent of households that simultaneously have young children, use center-based care rather than family or friends, and pay market rates rather than subsidized rates. Reporting childcare price growth as if it applies to all American families is innumerate.
You would distinguish stocks from flows. The 81 percent home-price increase since 2017 describes the situation of would-be first-time buyers, who are 5 to 10 percent of households per year. Existing homeowners—about 65 percent of households—are locked into sub-4 percent mortgage rates and have seen their housing costs fall in real terms. These are opposite stories, not the same story.
You would distinguish severity from prevalence. A statistic that includes the destitute tail and the discretionary grumblers in the same number tells you nothing about either.
You would distinguish self-inflicted costs from imposed ones. Someone who chose to take on $200,000 of debt for a degree with predictable low earnings is in a different situation from someone who was actively defrauded. Lumping them together as “student-debt-burdened” insults both.
You would examine the policy counterfactual. What would tuition look like today if federal loan limits had been frozen in 1985? What would coastal housing look like if zoning had been loosened 20 years ago? You don’t have to know the answers with precision, but you do have to acknowledge that the questions exist.
And you would compare across categories that both have and haven’t been heavily subsidized. Why is a flat-screen TV affordable and a college education not? Why is a smartphone affordable and a hospital stay not? The answer is not a coincidence, and it is not because flat-screen TVs are easier to make than college degrees. It is because Congress has spent 60 years trying to make college affordable and has spent zero years trying to make TVs affordable.
In closing.
The affordability crisis is real. A meaningful share of Americans are in genuine catastrophic distress and need help. A larger share of talented young Americans are being squeezed out of the lives their parents had. Both groups deserve serious attention. Neither is well served by being lumped into a single statistic alongside people who are mad about streaming bundles and $16 salads. Holding both of those positions at once is hard but indispensable.
The lumping-in allows the affordability discourse to keep recommending the same policy menu that has, in category after category, made the underlying problem worse. It is what allows both political parties to claim they are addressing a crisis while pursuing policies aimed at different populations. It is what allows everyone to skip the awkward question of why, after 60 years of subsidies, the things we subsidized got more expensive while the things we didn’t got cheaper.
If you want to take affordability seriously, you have to break the number apart. You have to ask which problem you’re solving. You have to ask what we’ve tried before. You have to ask why what we tried didn’t work. And you have to be willing to hear answers that don’t flatter your priors, because none of the answers in this territory flatter anyone—including, on more than one count, us.
Until then, the next person who tells you that half of Americans can’t afford to live is telling you a wrong number. They may not know it. The number itself does.