Income vs. Wealth Inequality: What Every American Should Know

Last updated 3 days ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.

When Barbara Hauke moved into her Jersey City, New Jersey, condo in 2004, her annual property taxes were about $5,000. She is retired now, and her taxes have climbed to roughly $20,000 a year. Her income did not quadruple. The thing she owns got more valuable, and the bill on it comes due in cash.

Hauke is standing on two ladders at once, and that is where nearly everyone gets lost in the inequality numbers. Income is what arrives this year. Wealth is what has piled up since you started, minus what you owe.

The wealth gap looks worse than the income gap because wealth is a stock that compounds and is taxed only when something is sold, while income is a flow that taxes and transfers squeeze every year, so a household can sit high on one ladder and near the bottom of the other without having done anything wrong.

The distance between the two is not subtle. In 2024, the top 5 percent of households received 23.1 percent of aggregate household income. In the first quarter of 2026, the top 0.1 percent held 14.4 percent of all household net worth. Five percent of households against one tenth of one percent of households.

Two ladders, and only one has a rung with your name on it

The Census Bureau publishes the income ladder in dollars. In 2024, households in the lowest quintile had incomes of $34,510 or less, the second fifth topped out at $65,100, the third at $105,500 and the fourth at $175,700, and the top 5 percent of households began at $335,700. The same report puts the median at $83,730 and the 90th percentile at $251,000.

Find your household on that list and you have your rank, to the dollar. Now try the same thing with what you own.

The Federal Reserve’s Survey of Consumer Finances (SCF) is the official wealth survey. In its 2022 wave, published in October 2023 and still the most recent, real median net worth surged 37 percent to $192,900. That is the middle of the country. What it publishes by rank is income rank, not wealth rank: the median net worth of families in the top decile of usual income was $2,556,200.

So there is no official answer to the question “what net worth puts me in the top 10 percent.” The Fed’s quarterly Distributional Financial Accounts (DFA) offer interactive charts and tables that let users explore the level, composition, and share of U.S.

household wealth held by five percentile groups, but you have to already know which group you are in. That asymmetry is not an oversight. Income is a number a survey can ask for and a pay stub can confirm; net worth is a number somebody has to build, line by line, out of things nobody has recently priced.

A flow you can count, a stock somebody has to appraise

Census money income is pretax and does not account for the value of in-kind transfers. Read that plainly: the figure you just looked up ignores your tax bill, ignores refundable credits, ignores food and housing assistance, and ignores capital gains. It is a deliberately narrow ruler, and it sits behind most headline income statistics you will ever see.

Net worth is the difference between families’ assets and liabilities. Your house counts, at what it would sell for, and the mortgage counts against it. Your 401(k) counts too: 54.3 percent of families held a retirement account in 2022, up from 50.5 percent in 2019, and among families that had one the median balance was $86,900 while the mean was $334,000.

Social Security does not count, and neither does a traditional pension you have not started drawing. Measured net worth counts assets you own, and a claim on a future benefit stream is not a thing you can sell.

A January 2026 Federal Reserve working paper by Landel, Love and Smith builds an alternative that augments traditional net worth with the actuarial present values of expected future payment streams from labor-market earnings, Social Security, defined-benefit pensions, annuities, life insurance and government transfers. Adding those claims materially lowers measured concentration, which is one of the two or three real fights in this field.

Somebody has to sit in a living room and assemble all of it. The SCF is fielded by trained interviewers, and in the 2022 SCF, the median interview length was about 110 minutes, up from 100 minutes in 2019, with the hardest cases running well past three hours. Everything in the paragraph above rests on a few thousand of those conversations.

Why the stock is always more concentrated than the flow

Set the two distributions side by side and the shapes do not match.

The categories do not line up, and that is the point rather than a flaw in the table. Income shares are cut by income rank; wealth shares are cut by wealth rank; the two rankings sort the same country differently.

Emmanuel Saez and Gabriel Zucman give the mechanism in a single sentence: the increase in wealth inequality in recent decades is due to the upsurge of top incomes combined with an increase in saving rate inequality. Read it slowly, because it explains the whole article. The flow feeds the stock. The households with the biggest flow also turn the largest share of it into assets, and they do it again the next year, on top of what last year left them.

At the other end the same arithmetic runs backwards. In the first quarter of 2026 the bottom half of households held 2.5 percent of total net worth while carrying 30.4 percent of all household liabilities, including 51.8 percent of consumer credit. A household with a car loan, a card balance and no equity is not sitting at zero. It is below zero, and its interest compounds in the same direction the top’s returns do.

The middle’s balance sheet is a house; the top’s is a company

Households in what the Fed calls the next 40 percent, the 50th to 90th percentile of wealth held 46.5 percent of the country’s real estate and 11.6 percent of its corporate equities and mutual fund shares in the first quarter of 2026. The top 0.1 percent held it the other way round, with 24.2 percent of all corporate equities and 28.3 percent of all unincorporated business equity, against 4 percent of the real estate.

That is not a difference of degree. The middle owns one leveraged, illiquid, locally priced asset that it also sleeps in. The top owns claims on companies. Almost everything else here follows from that contrast, because what your wealth is made of decides how it grows, how it is taxed, and what happens to it when the market turns.

When the market turned, the difference showed. Between 2007 and 2010, median net worth fell 38.8 percent, and the mean fell 14.7 percent. A house bought with a mortgage is a leveraged bet, and the debt does not shrink when the price does.

Sheila Ramos was 58 and had custody of three grandchildren when she fell behind on the mortgage. They had flown from Florida, after she fell hopelessly behind on the payments for her three-bedroom home, to a family-owned patch of rural land on Hawaii’s Big Island.

On a July night in 2009 they pitched a tent there, with no electricity, and started a new life. Nearly three years on they were still in a semi-permanent encampment, and she was scavenging the local dump for salvage. Her income loss and her wealth loss were one event, which is exactly what makes a house different from a share portfolio.

A house is also worth whatever one appraiser says it is worth. In 2019, the home was appraised at over $1.4 million, Paul Austin and Tenisha Tate-Austin said of their Marin City house, but one year later it came back at just $995,000.

They took down the family photographs and the other signs of who lived there, asked a white friend to pose as the owner, and the next appraisal came in far higher. They sued, and settled the federal housing discrimination case for an undisclosed sum. Nobody re-prices an index fund according to who is holding it.

Why two federal agencies publish different answers

The SCF goes looking for rich households on purpose. Alongside a national area-probability sample it draws what it calls a “list sample,” this group is drawn from a list of statistical records derived from tax returns. And then it stops short of the summit: persons listed by Forbes magazine as being among the wealthiest 400 people in the U.S. are excluded from sampling.

Read that twice. The government’s flagship wealth survey deliberately over-samples the rich for accuracy, then deliberately leaves out the 400 richest Americans for confidentiality. Any claim about the very top built on the SCF alone is built on a sample that omits the very top.

The quarterly wealth shares are stitched together from two worlds. The Fed builds them in three steps: generate a balance sheet from the SCF that is conceptually consistent with the components of aggregate household net worth in the Financial Accounts, then interpolate and forecast that reconciled balance sheet into quarters nobody surveyed, then scale the result so it sums to the published aggregate.

The wealth figure you read for early 2026 is therefore an estimate resting on interviews conducted in 2022, and the Census income figure for 2024 is a separate survey with a different concept of what counts. Two agencies, two instruments, two answers, and neither is wrong.

What April actually moves

Money income is measured before the government does most of what it does about income. The Census Bureau’s Supplemental Poverty Measure shows the difference: refundable tax credits, SNAP and housing subsidies are only included in the SPM, and necessary expenses such as taxes, medical costs and work-related costs are deducted from its resources but ignored by the official poverty measure. Same households, different ruler, different answer.

Gerald Auten and David Splinter argue that once you do this properly, the income story changes a great deal.

Measuring income on a national-income basis and counting taxes and transfers, they report that increases in income shares for the top 10 percent and top 0.1 percent are smaller than Piketty, Saez and Zucman find, that after-tax shares for those groups are little changed, and that the top 1 percent’s share rose about 4 percentage points between 1988 and 2012 against the Congressional Budget Office’s 9-point rise in the before-tax share from 1979 to 2019.

That is the strongest form of the case that the income gap is overstated, and it is not a fringe position. It turns on measurement choices anyone can inspect and argue with.

The federal income tax is only part of your tax bill, though, and the rest of it leans the other way. The Institute on Taxation and Economic Policy (ITEP), which argues for more progressive taxation, finds in the seventh edition of its Who Pays?

study that the nationwide average effective state and local tax rate is 11.4 percent for the lowest-income 20 percent, 10.5 percent for the middle 20 percent and 7.2 percent for the top 1 percent. Barbara Hauke’s $20,000 property tax bill is a state and local tax, and it is charged on the value of her condo rather than on what she earns. That is ITEP’s whole point in one retiree.

Why the same code barely touches the stock

One sentence in the Internal Revenue Code separates the two ladders. Gain is the excess of the amount realized therefrom over the adjusted basis, computed on the sale or other disposition of property. No sale, no amount realized, no gain, no tax. Your house can double and you owe nothing until you sell, and a founder’s stake in a company works the same way.

When a sale does happen, the rate is gentler than the rate on wages. For 2025, the tax rate on most net capital gain is no higher than 15 percent for most individuals, with a 0 percent rate up to $48,350 of taxable income for a single filer and $96,700 for a married couple filing jointly, while wages run up the ordinary brackets.

The sale can also be skipped entirely. When an owner dies, the basis of inherited property becomes the fair market value of the property at the date of the decedent’s death, so a lifetime of unrealized appreciation is erased for income tax purposes. The estate tax is the backstop, and it is a narrow one: the One Big Beautiful Bill Act increased the basic exclusion amount to $15,000,000 for calendar year 2026.

Stack those rules and you get an asset that grows untaxed, is taxed lightly if sold, and is never taxed as income if held until death. Nobody snuck that in. It is the ordinary operation of a realization-based income tax, and it is the reason the machinery that compresses the income distribution leaves the wealth distribution close to where it found it.

Leaked IRS files reported by ProPublica put faces on the arithmetic. In 2007, Jeff Bezos, then a multibillionaire and now the world’s richest man, did not pay a penny in federal income taxes, and he achieved the feat again in 2011. Elon Musk owed nothing in 2018.

Peter Thiel’s case shows the same rule running through an account built for ordinary savers. In January 1999 he bought his PayPal founders’ shares inside a Roth individual retirement account, and Thiel paid $0.001 per share, a tenth of a penny, for 1.7 million shares, a total outlay of just $1,700 against an annual Roth contribution cap of $2,000. Gains inside a Roth are not taxed when they come out. ProPublica valued the stake that grew from that $1,700 at $5 billion, tax free.

ProPublica also put a number on the pattern. Forbes estimated the 25 richest Americans’ worth rose a collective $401 billion from 2014 to 2018; they paid $13.6 billion in federal income taxes over those five years, which the outlet called a true tax rate of only 3.4 percent.

That figure carries a serious objection, and it is worth putting in the words of the people who make it. The denominator is not income under any definition the tax code recognizes; it is the change in an outside magazine’s estimate of what somebody’s assets would fetch. Lawyers at the Competitive Enterprise Institute, arguing the point to the Supreme Court, wrote that Eisner v.

Macomber makes perfectly clear that the lynchpin for Sixteenth Amendment incomes is realization by the taxpayer. On that view a rate computed against unrealized appreciation is not a tax rate at all, because there is no income underneath it. Both things can be true: the comparison is arresting, and it is not measuring what a tax rate normally measures.

The files also left the IRS unlawfully. Charles Littlejohn, 38, a contractor, pleaded guilty to unauthorized disclosure of tax returns and return information, and at his sentencing federal District Judge Ana Reyes called his conduct “an attack on our constitutional democracy” and gave him the maximum five years. She said he had targeted a sitting president and that it could not be open season on elected officials.

Whether the stock can be taxed at all

Charles and Kathleen Moore invested in the American-controlled foreign corporation KisanKraft, which supplied equipment to small-scale Indian farmers. From 2006 to 2017 the company generated a great deal of income but did not distribute it to its American shareholders. The Mandatory Repatriation Tax then billed the Moores $14,729 on their share of earnings they had never seen. They paid it, sued for a refund, and argued that taxing money never received is an unapportioned direct tax the Constitution forbids.

They lost, and the terms of the loss matter more than the result. The Court upheld the tax on the ground that Congress may attribute a company’s realized but undistributed income to its shareholders. It set aside the Moores’ reliance on Eisner v. Macomber, where the question was whether a distribution of additional stock to all existing shareholders was taxable income and the Court said no, that income requires realization. It expressly declined to decide whether a tax on unrealized appreciation would be allowed.

Myrtle Macomber is the older half of that story. She received a pro rata stock dividend on her Standard Oil Company of California shares and the government taxed it as income, and the Court held that Congress was not empowered by the Sixteenth Amendment to tax, as income of the stockholder, without apportionment, a stock dividend made lawfully and in good faith. More than a century later, both sides of the wealth-tax argument still cite her case.

States answer to different constitutions and have gone further. Chris Quinn and other Washington residents challenged the 7 percent tax the legislature enacted in 2021 on the sale or exchange of certain long-term capital assets, and on March 24, 2023, the Washington Supreme Court held that the capital gains tax is an excise tax under Washington law, which placed it outside the state constitution’s limits on taxing property. A tax that the state calls an excise on a transaction survives where a tax on holdings might not.

Thirty-five, well paid, and worth less than nothing

Jenni and Sean Gritters moved to Seattle after bachelor’s and master’s degrees in Boston. The Gritters owe about $125,000 on four student loans, which leaves them with a negative net worth of $93,500. On the income ladder they are doing fine. On the wealth ladder they are below the bottom rung, and the first fact is part of what caused the second.

Education loans subtract from net worth like any other debt. 21.8 percent of families held education loans in 2022, with a median balance among borrowers of $24,500, and the Bulletin notes that the debt is concentrated among higher-income families, more than half of the balance sitting with the top two income quintiles. That reads oddly until you remember what the degree bought.

Age explains most of the rest of the gap. In 2024, householders aged 45 to 54 had the highest median income at $116,800, followed by householders 35 to 44 at $106,100, while householders 65 and over had the lowest at $56,680. Median family net worth in 2022 climbed the other way for the young: $39,000 under 35, $135,600 at 35 to 44 and $247,200 at 45 to 54.

Income peaks in your late forties and early fifties and then falls. Wealth keeps climbing past that, because it is the leftover of every year of income you did not spend. A 35-year-old with a strong salary and a negative net worth is not failing; they are early. The comparison that would tell them something useful is against other 35-year-olds, and a national percentile does not make it.

Did it widen in your lifetime, or does it just feel that way

Both series say it widened, and they have been saying so for decades. The Census Gini index runs from 0.0 for perfect equality to 1.0 for total inequality, and the Gini index was 0.488 in 2024, not statistically different from 2023. Go back to the start of the series and the household Gini was 0.386 in 1968. It moves over decades, not over years, which is why one year’s headline almost never means anything.

The wealth series moved further over a shorter window. When the DFA begins, in 1989:Q3 the top 0.1 percent held 8.6 percent of household net worth and the top 1 percent held 22.8 percent; by the first quarter of 2026 those were 14.4 and 31.6 percent, a rise of roughly nine percentage points for the top 1 percent. Over the same span the bottom 50 percent went from 3.5 percent to 2.5 percent of net worth.

Nine points of a country’s net worth is a large amount of ground to change hands, and it happened on the Federal Reserve’s own numbers rather than an advocacy group’s. That part is not a feeling.

What the economists are actually fighting about

Two disagreements keep getting mistaken for one. The first is about measurement, and it is settleable in principle. The second is about what a tax ought to reach, and it is not.

On measurement, Saez and Zucman argue the surveys cannot see the top and go to tax records instead. Capitalizing income tax data, they find that the share of wealth owned by the top 1 percent families reached 42 percent in 2012, with the top 0.1 percent’s share growing from 7 percent in 1978 to 22 percent in 2012. That sits well above the Federal Reserve’s estimate, and the gap is method rather than arithmetic: one team scales up income flows into implied assets, the other asks families what they own.

Auten, Splinter and scholars at the American Enterprise Institute (AEI) push the opposite way. AEI’s case is that income is only a partial measure of prosperity, and that studying changes in the distribution of consumption and expenditure helps amplify this picture, since income is valued largely because it allows consumption. The Federal Reserve’s own comprehensive-wealth work makes the parallel point about the stock: count pension and Social Security claims as the assets they functionally are, and measured concentration drops.

Both camps are making real arguments, and they fail in different ways. Saez and Zucman need their capitalization rates to be right. Auten and Splinter need their allocation of unreported income and untaxed transfers to be right. Neither has a ledger of what every American owns, because no such ledger exists, and a reader who wants one number that everybody accepts is going to be disappointed.

The value disagreement sits underneath and does not budge when the measurements improve. Hall and Rabushka would swap the income base for a consumption base: the individual wage tax would be imposed on wages (and salaries) and pension receipts, paired with a business tax, a combination the Congressional Research Service (CRS) says can be viewed as a modified value-added tax (VAT).

CRS also notes that a consumption base is neither inherently superior nor inferior to an income base. That is a choice about what a tax should touch, and no distributional table decides it for you.

What a percentile cannot tell you about your household

Every figure here is an estimate with a range around it. A margin of error, when added to and subtracted from the estimate, forms the 90 percent confidence interval: the 2024 median of $83,730 carried one of $1,050, and the 90th percentile of $251,000 carried $2,731. Census marks a year-to-year change only when it can be distinguished from zero, and the 2023 to 2024 change in the median could not.

Your own number is shakier than the estimate. In the Federal Reserve’s 2025 survey of household economics and decisionmaking, 30 percent had income that varied at least occasionally through the year, up from 28 percent in 2023, and 58 percent of the self-employed said their income varied month to month. A percentile describes a year. Plenty of households do not have a typical one.

So here is what you can honestly take away. You can find your income rank in the Census tables to the dollar. You can find which of the Fed’s five wealth bands you fall into, but not a precise wealth rank, because no agency publishes one. And you should expect the two ranks to disagree, most sharply if you are young, if you rent, or if the degree that raised your income has not yet paid for itself.

The distance between your two ranks is mostly a fact about what you own and how long you have owned it, not a verdict on how you have done. Wealth is time plus income minus spending, and the youngest households have had the least of the first. That is the one part of this subject nobody is fighting about.