Cheap Things,
Expensive Lifewhy a richer country feels poorer

Here is the version you've heard, and it is not a strawman, because it is substantially correct.

Since the late 1980s, the share of American wealth held by the top one percent has climbed from a little under a quarter to just under a third. The top tenth of one percent — around 130,000 households — went from holding 8.6% of everything to 14.5%. The bottom half of the country, some 66 million households, holds about 2.5%. At the depths of the post-2008 wreckage, that number touched 0.4%. Not four percent. Four tenths of one percent.

Put those two facts side by side and you get the line that does the rounds: the top 1% now owns more than the bottom 90% combined. It's true. It's from the Federal Reserve. It is not a talking point invented by an activist with a grudge; it's the central bank's own quarterly accounting of who owns America.

From there the conclusion writes itself, and you've seen it a thousand times. The rich are eating the country. The middle class is being hollowed out. Wealth is flowing upward and it is flowing out of everyone else's pockets. The pie is being redivided, and you are getting less pie.

That last step is where things go wrong.

Not because the concern is misplaced — it isn't, and by the end of this piece I'll argue the pessimists are onto something more serious than the thing they're describing. It goes wrong because a share is a ratio, and a ratio tells you nothing whatsoever about the size of the thing being divided. Over the same period that the top 1% grew its slice by nine percentage points, the pie itself roughly quadrupled. Both facts are true. They are describing the same country. And they point in opposite directions.

So this essay does three things in sequence, and I want to flag the shape up front so it doesn't feel like a bait-and-switch when the ground moves under you.

• The shares

  The pessimists are right, and the data is worse than most of them realise — though not in the place they think.

• The levels

  In absolute, inflation-adjusted dollars, essentially every group in America is richer than its counterpart three decades ago. Not marginally. Substantially. This is not a fringe claim; it's the Congressional Budget Office's.

• The resolution

  Both of those things are true, both are measured accurately, and the reason they feel irreconcilable is that we have been measuring prosperity in one basket of goods and living it in a completely different one. The things that got cheap are the things you enjoy. The things that got expensive are the things that determine where your children end up.

Four ideasbefore the charts start lying to you

This section is short and it is load-bearing. Skip it and the rest of the piece will seem to contradict itself.

Average and median are not the same thing, and the gap is the point

Imagine ten people in a bar. Nine of them earn $50,000. The tenth earns $50,000 too. The average income is $50,000; the median — the person standing in the middle if you line everyone up — is also $50,000. The two measures agree because the group is uniform.

Now Jeff Bezos walks in.

The average income in that bar is now something north of $20 billion. The median has not moved by a single dollar. It is still $50,000, because the person in the middle of the line is still the same guy, and Bezos just stands at the end of it.

That's the whole idea, and it matters enormously, because almost every cheerful statistic about the American economy is an average and almost every gloomy one is a median. "GDP per capita is at a record high" is a bar with Bezos in it. "The typical worker" is the guy in the middle of the line.

The Social Security Administration publishes both for American wages, and the gap between them is one of the most quietly devastating series in the data. In 1991 the median American wage was 72% of the average. By 2014 it had fallen to 64.7%. Same country, same workers, and the middle of the line had drifted a long way from the arithmetic centre of gravity — which is exactly what you'd expect if the gains were piling up at one end. Hold that thought; the series does something unexpected after 2014.

A slice is not a quantity

The second idea is the one this entire essay turns on, and it's embarrassingly simple.

If your share of a pizza falls from 20% to 15%, you might have less pizza. You might also have more pizza, if someone brought out a much bigger pizza. The share alone genuinely cannot tell you. You need to know the size of the pizza, and nobody arguing about inequality on the internet ever mentions the size of the pizza.

Every chart in this piece is one of two kinds. Share charts — percentages, ratios, slices. Level charts — actual inflation-adjusted dollars, quantities of pie. They answer different questions, they frequently point in opposite directions, and conflating them is the single most common error in this debate, committed enthusiastically by both sides.

Who is "the bottom half," exactly?

The buckets

To keep this navigable, five groups, same colours, every chart. They match the Federal Reserve's Distributional Financial Accounts, which is the backbone source here. Where a chart uses different groupings — the CBO's income work uses quintiles, ranked by income rather than wealth — I've tagged it so you know the ground has shifted.

One last piece of throat-clearing. Economists argue fiercely about how much wealth and income have concentrated in America; the estimates from different research teams vary by more than you'd think possible given they're using the same tax data. They do not argue about the direction. Everything here leans on the government's own series — the Federal Reserve, the Congressional Budget Office, the Census Bureau, the Bureau of Labor Statistics — precisely to stay out of that fight. Notes at the end for anyone who wants it.

The pessimists are rightand the surprise is who lost most

Start with the Federal Reserve's Distributional Financial Accounts, which since 1989 has published a quarterly estimate of who owns America's household wealth. Here is what happened between 1989 and the end of 2025.

The tenth of a percent went from 8.6% of all household wealth to 14.5%. Their share grew by about two-thirds. The one percent as a whole went from 22.8% to 31.9%. The professional class — the 90th to 99th percentile, successful doctors, senior engineers, people with a paid-off house and a healthy 401(k) — went from about 38% to 36.4%. They lost ground. The middle went from 35.7% to 29.4%. The bottom half sits at roughly 2.5%, having bottomed out at 0.4% after 2008.

Sit with the third and fourth numbers, because they're the ones that don't get quoted. In percentage-point terms, the biggest loser over the last three and a half decades was not the bottom half. It was the middle. The bottom half never had much to lose. The professional class lost about 1.6 points. The middle lost about 6.3.

The mechanism, and the wages underneath it

None of this is mysterious. Wealth in the middle of the distribution is overwhelmingly the house; wealth at the top is overwhelmingly equity. Financial assets have massively outperformed residential real estate for three decades, so every bull market widens the gap mechanically, without anyone taking anything from anyone. It isn't a conspiracy. It's what happens when one group's savings sit in an appreciating financial instrument and another's sit in a depreciating roof over a slowly appreciating patch of dirt.

Wealth is a stock; wages are a flow, and the flow tells the same story with less noise.

So: on shares, the pessimists win, and they win comfortably. Wealth concentrated. The concentration is real, it's large, it's documented by the central bank, and it has been going in one direction for three and a half decades.

Now turn the chart around.

Everybody got richerand it isn't close

Same country. Same period. Absolute dollars, adjusted for inflation.

The Congressional Budget Office's Trends in the Distribution of Family Wealth, 1989 to 2022 is the most careful public accounting of this that exists, and its headline number should be in every one of these arguments and never is: total American family wealth went from $52 trillion to $199 trillion, in constant 2022 dollars. It almost quadrupled.

That is a real, inflation-stripped quadrupling in thirty-three years — roughly 4% a year, compounding. And it did not go only to the top.

Look at individual points in the distribution and the pattern holds. Between 1989 and 2022, family wealth rose 232% at the 25th percentile, 131% at the median, and 148% at both the 75th and 90th. The single fastest-growing point in the entire American wealth distribution was the household one-quarter of the way up from the bottom. Median family wealth in 2022, on the CBO's measure: $504,000.

Income tells the same story

The CBO's parallel work on income, running back to 1979 and measured after taxes and government transfers — the number that actually determines what a household can spend — shows the bottom quintile roughly doubling, the middle three quintiles up about 65%, the top quintile up 144%, and the top 0.01% up more than sevenfold. Every group up. Enormously unevenly, but every group up. And the Census Bureau's median household income hit $83,730 in 2024, an all-time record in inflation-adjusted terms.

Now let me undercut all of that

Two caveats, and I'm putting them here rather than in a footnote because if I don't, you'll find them yourself and stop trusting me. A third, briefly: 2022 sits at the top of an extraordinary asset boom, and a different endpoint tells a different story.

And one thing that runs the other way

Having spent three paragraphs undermining the optimistic case, here's something that undermines the pessimistic one, and almost nobody has noticed it. The concentration trend reversed after 2019.

Go back to that SSA median-to-average wage ratio. It fell for a quarter century to 64.7% in 2014 — and then it turned around. By 2023 it was back to 67.6%, erasing roughly a decade of drift.

The labour economics matches. Work by David Autor, Arindrajit Dube and Annie McGrew documented what they called an "unexpected compression": between 2019 and 2024, real wages at the 10th percentile rose about 15% — more than that group's cumulative gain over the preceding forty years — against roughly 6% at the median and 7% at the 90th percentile. For the first time in four decades the bottom of the American wage distribution grew fastest, reversing by their estimate something between a quarter and 40% of the entire four-decade rise in 90/10 wage inequality.

Whether that survives a slacker labour market is an open question. But it demonstrates something important: the trend is not a law of physics. It reversed, hard, within living memory, under identifiable conditions. Any account that treats rising inequality as an inexorable feature of modern capitalism has to explain 2019–2024.

The basket split in two

Here is where the essay earns its keep.

We have established two things that seem to be in tension. Wealth concentrated dramatically at the top. Also, essentially everyone got richer in real terms. Both from government data, both robust.

And yet: ask an American whether life is easier than it was for their parents and you will not get a happy answer. Consumer sentiment has been miserable through periods of record employment and record real median income. Something in the lived experience is not being captured by "your inflation-adjusted income is at an all-time high."

It isn't sentiment, or nostalgia, or media pessimism. It's a measurement problem, and it's this: inflation is an average, and the average conceals a split so violent it makes the aggregate almost meaningless.

Look at those two groups and ask what actually separates them.

A television is a thing you enjoy. A hospital is a thing that determines whether you live. A toy is a thing your child plays with. A university is a thing that determines what your child's life looks like at forty. Childcare is what determines whether both parents can work. Housing is not merely shelter — in America the house determines the school, the school determines the peer group, the peer group and the school determine the university, and the whole chain determines where your children land in the distribution we spent Section III measuring.

The goods that collapsed in price are the goods that make life pleasant. The goods that exploded in price are the goods that make life mobile.

And this is precisely why the aggregate statistics and the felt experience diverge. When the CBO says median family resources rose, it is telling the truth using a basket in which cheap televisions and cheap clothing count. When a 35-year-old says they can't do what their parents did, they are also telling the truth, about a basket containing a house near a decent school, a university education for two children, and health insurance that doesn't collapse under a serious diagnosis.

Both people are correctly reading their instruments. The instruments are pointed at different things.

The house is the clearest case

Housing deserves its own section because it's where the abstraction becomes concrete and personal.

Where shares stop being about fairness

Everything so far could still be filed under "unequal but improving." Section III says the distribution got lopsided; Section IV says everyone's absolute position improved anyway; Section V says the improvement was concentrated in things that matter less than the things that got worse.

This section closes the loop, and it's the most important research in the piece.

Raj Chetty and colleagues asked the simplest possible question about the American Dream: what fraction of children grow up to earn more than their parents did? Not more than the average, not a bigger share — just more than their own mother and father, adjusted for inflation. It's about as clean a definition of generational progress as you can construct.

Now the finding that welds this essay together. Chetty's team decomposed the decline to ask what caused it: was it that the economy grew more slowly after 1980, or that the growth was distributed differently? They ran both counterfactuals. Faster growth alone doesn't come close to restoring the old rates. Distributing the growth America actually had the way America distributed growth in 1940 would reverse more than 70% of the decline.

That result answers the question this essay has been circling. Throughout Sections III and IV I've been implying that shares are a fairness question and levels are a welfare question — that concentration is about justice while absolute growth is about whether people are okay. Chetty demonstrates that's wrong. The distribution of growth mechanically determines whether people beat their parents. It isn't a moral overlay on the numbers; it's an input to the most basic measure of whether a society is working.

Which is also why "but everyone's absolutely richer" fails as a rebuttal, even though it's factually correct. Nobody experiences their life as a comparison to a 1989 statistical abstraction. People experience their life as a comparison to their parents, their siblings, their neighbours, and the version of themselves they expected to be. On that measure — the one people actually use — half the country is losing, and it's the middle that's losing hardest.

The test that could have sunk this argument

Everything so far is internal to America — comparing 2022 America to 1989 America. That's vulnerable to a specific objection: maybe this is just what advanced economies do. Maybe rich countries inevitably see healthcare and education costs balloon, housing tighten, mobility fall. Maybe there's no counterfactual.

There is a counterfactual, and it's other rich countries.

The trend line is worse than the level. Across the 2010s — a decade of expansion, record markets, and no pandemic — American life expectancy improved by 0.26 years in total. Over the preceding five decades, the average gain was 1.78 years per decade. Nearly a sevenfold slowdown, in a country getting demonstrably richer the entire time, while its peers kept improving.

So: the richest large country on earth, in the middle of the greatest accumulation of household wealth in human history, converted that wealth into less additional life than countries spending half as much per head.

If America had converted its growth into wellbeing at the rate its peers managed, this essay would have a much weaker thesis. It didn't. That's the test, and the thesis survives it.

Three claims, and I'd defend all of them.

The pessimists are right about the shares — and they understate the middle. The group that gave up the most ground wasn't the poor; it was the 50th-to-90th percentile, the people who'd call themselves comfortably middle class.

The optimists are right about the levels, with an asterisk. Real resources rose for essentially every group and the pie roughly quadrupled — but a large share of that is Social Security wealth you can't spend, and a population simply older and better-credentialed than it used to be.

And both are answering a question that misses the point. Measured in televisions, clothing, appliances and software, Americans are staggeringly better off. Measured in houses near good schools, university degrees, childcare, and medical care that doesn't bankrupt you, they are worse off — and those are precisely the goods that determine where a family's children end up.

That's the resolution. We measure prosperity in a basket that got cheap and we live it in a basket that got expensive. The felt sense of decline in a country of rising real incomes isn't false consciousness or media-induced gloom. It's an accurate reading of a different and more consequential dataset — the one where the coin flip on whether your kids do better than you has replaced the near-certainty your grandparents had.

The thing worth arguing about, then, isn't whether the rich got richer. They did, it's measurable, and the argument is over. It's why the four things a family needs most — a home, a doctor, a degree, and someone to watch the children — are the four things that consistently outran everyone's ability to pay for them, in the richest country that has ever existed.

That's a narrower question than "is capitalism working." It's also one that has answers, and different answers in different places, which is a great deal more than can be said for the argument we're currently having.