Time: Millionaires of 1955

The Million Dollar Question: In 1955 the top federal income tax rate was 91%. The 263 taxable returns that reported adjusted gross income of $1 million or more paid federal income tax equal to roughly what share of that income?
A) About 91% B) About 75% C) About 53% D) About 25%

Read on for the answer.

Every argument about taxing the rich eventually reaches for the same number: ninety-one per cent, the top federal income tax rate through the whole of the Eisenhower administration. Depending on who is holding it, the number proves either that the country once taxed great fortunes hard and prospered anyway, or that it only pretended to.

Both camps tend to quote the rate and stop there. This instalment of the series does what the 1900 and 1925 pieces did: it goes to the federal government’s own count of who reported a million dollars, how they earned it, and what they actually sent to the Treasury. The answer is neither 91% nor nothing, and the gap between the two is the most useful thing the year has to teach.

The rate everyone quotes

The schedule itself is not in dispute. The Internal Revenue Service’s instructions for the 1955 Form 1040 print it in full. For a single filer, taxable income was taxed at 20% on the first $2,000 and climbed through twenty-four brackets to “$156,820, plus 91%” of everything over $200,000. Rates had already reached 50% at $16,000 of taxable income and 75% at $50,000.

Married couples filing jointly got a separate schedule, and it was not the same table with bigger numbers. Since 1948 the law had let a couple split their combined income down the middle and pay twice the tax a single person would owe on half of it. The instructions call this the “split income” method. The practical effect was that the 91% bracket began at $400,000 for a couple, and every threshold below it doubled too. A married man with $95,000 of taxable income faced a marginal rate of 72%. A single man with the same income faced 87%.

Two more features of the statute took the edge off before any planning began. The 1954 code capped the whole computation: section 1(c) provided that the tax “shall in no event exceed 87 percent of the taxable income for the taxable year,” a limit the Treasury’s regulation at 26 CFR 1.1-2 still records. And long-term capital gains, on assets held more than six months, did not go through the schedule at the top rate at all. As the Statistics of Income for 1955 explains, only half of the excess of net long-term gain over net short-term loss was included in adjusted gross income, and an “alternative tax” capped the rate on that whole excess gain at 25%. The IRS’s own statisticians noted that the alternative tax became the better deal once taxable income reached $18,000 on a separate return or $36,000 on a joint one, which is to say for almost everyone with real money.

So the 91% bracket existed, applied to ordinary income above $200,000 or $400,000, and could never take more than 87% of taxable income overall. A ferocious rate, and a narrower one than the headline suggests.

The 263 returns

The best evidence of what it did is the IRS’s own tabulation. Statistics of Income for 1955, published in 1958, sorts every individual return filed for the year by adjusted gross income, and the top class is “$1,000,000 or more.”

There were 267 of them. Out of 58,250,188 individual returns filed in the United States, 267 reported adjusted gross income of a million dollars or more — about one return in 218,000. Together they reported $567.6 million of AGI, roughly 0.23% of the national total of $249.4 billion. Four of the 267 owed no income tax at all.

The other 263 did, and Table 8 of the report sets out what they paid. Their combined AGI was $550.9 million and their taxable income $452.7 million. Their income tax after credits came to $291.0 million — an average of $1,106,410 per return.

That is 52.8% of their adjusted gross income, and the answer to the Million Dollar Question is C. The report’s own “effective tax rate” column, which divides tax by taxable income rather than AGI, shows 64.3%. Either way, the 263 largest taxable incomes in the country paid more than half of what the tax code counted as their income. Collectively they supplied just under 1% of all individual income tax collected that year, $291 million out of $29.6 billion.

The same table splits them by method, and the split is the whole story in miniature. Only 31 of the 263 paid tax at the regular rates. Their combined tax came to 82.9% of their combined taxable income, pressed close to the 87% ceiling. The other 232 used the alternative capital gains computation, and their effective rate on taxable income was 62.1%.

The reason is visible in what they reported. Table 3 of the same report breaks down AGI by source for all 267 returns in the top class:

• Dividends: $286.2 million, about 50% of their AGI.
• Net gain from sales of capital assets: $248.1 million, about 44%.
• Salaries and wages: $7.8 million, about 1.4%.

A million-dollar income in 1955 was not a salary. It was a portfolio.

That capital gains line also hides something. The $248.1 million is the amount included in AGI, and only half of an excess long-term gain was included. Table 11 of the same report shows how much was left out. The 232 alternative-tax returns reported $491.2 million of net long-term gain at full value, against just $2.1 million of short-term gain, which was almost exactly offset by short-term losses; their net long-term gain in excess of short-term loss was $489.2 million. Half of that, about $244.6 million, never entered AGI. Add it back to the $550.9 million of AGI on the 263 taxable returns and the same $291.0 million in tax becomes about 37% (36.6%) of a broader income figure of roughly $795 million. That calculation is mine, not the IRS’s, and it measures only income that was realised and reported; it is still a long way from 91%.

The income, and the multiple

The earlier instalments asked what a million dollars held produced, and how many ordinary working lives that was worth. In 1955 the obvious benchmark is the corporate bond. Moody’s Aaa corporate yield averaged 3.05% across the twelve months of the year. A million dollars in top-grade corporates paid about $30,500 a year. The Social Security Administration’s national average wage index for 1955 was $3,301.44. Divide one by the other and you get 9.2.

In 1900 the same calculation gave 76. In 1925 it gave 34. By 1955 it was nine — and the coupon was taxed far harder than the untaxed coupon of 1900 or the roughly 9.5% bite of 1925.

Work the bill. Take a married couple whose only income is that $30,525 of bond interest. They take the standard deduction, which the 1955 instructions set at 10% of income up to $1,000, and two $600 exemptions. Taxable income is $28,325. On the joint schedule that is $8,520 plus 47% of the $325 over $28,000: $8,672.75, or 28.4% of their income. They keep about $21,850.

A single person with the same bonds gets one exemption and the single schedule, and owes $12,553.50 — 41.1%. Income splitting alone was worth nearly $3,900 a year on this one portfolio.

Now put the same million into municipal bonds. The Bond Buyer’s twenty-bond municipal index averaged 2.48% in 1955, which pays $24,783, and interest on state and local bonds was excluded from federal income altogether. The married couple ends up about $2,900 a year better off in munis than in Aaa corporates; the single investor, about $6,800 better off.

The municipal yield was 81% of the Aaa yield, so on these figures a tax-exempt bond beat a taxable one for anyone whose marginal rate was above about 19% — and the lowest bracket in 1955 was 20%. That compares index averages that differ in maturity, liquidity and credit, and I have not adjusted for any of that. But it suggests why a fortune that wanted income rather than growth had little reason to show up on a tax return at all.

The 2026 version of the same sum is less flattering to the millionaire. Moody’s Aaa averaged 5.88% in August 2026, so a million dollars in top-grade corporates pays about $58,800. Average weekly earnings for private-sector employees were $1,298.60 in August, or about $67,500 a year. That is a multiple of 0.87. If you prefer the ten-year Treasury, which closed at 5.26% on 29 September, it is 0.78.

Seventy-six, thirty-four, nine, less than one.

As for what $1 million in 1955 is “worth” today, the honest answer is that it depends on what you are measuring, and the spread is wide:

• Consumer prices: the BLS CPI-U averaged 26.8 in 1955 and stood at 334.98 in August 2026, a factor of 12.5. About $12.5 million.
• Wages: the national average wage index rose from $3,301 in 1955 to $69,847 in 2024, a factor of 21.2. About $21 million.
• Share of the economy: nominal GDP was $425.5 billion in 1955 and $30.86 trillion in 2025, a factor of 72.5. About $73 million.

The CPI answers “what could it buy?”, the wage index “how much labour could it command?”, and GDP “how large was it relative to everything else?” Measured the third way, a 1955 millionaire held something closer to a modern $70 million fortune — the moving-target problem in one line.

The ways around it

A rate that only reached ordinary income above $200,000 was always going to be met by people arranging not to have ordinary income above $200,000. The routes were not secret. Most were written into the statute.

Capital gains. The 25% alternative rate was the main one, and the SOI table above shows how thoroughly it was used: 232 of 263 top returns. Income that could be turned from dividends or salary into the appreciation of an asset paid at most 25%.

Leaving money inside a corporation. In 1955 corporations paid 30% on their first $25,000 of income and 52% above that. For an owner facing 91% personally, profit left in the company was taxed at 52% and could later come out as a capital gain at 25%, or not at all. This is the mechanism at the centre of the next section.

Municipal bonds. Excluded from federal income entirely, as the arithmetic above shows. A fortune held in municipals appeared on no line of any IRS table.

Oil. The percentage depletion allowance let producers deduct a fixed share of gross income from a well regardless of what the well had cost. Senator Edmund Muskie described it on the Senate floor in 1969, when it was still unchanged: the law “exempts from taxation 27½ percent of gross income for oil- and gas-producing properties, up to 50 percent of net income,” a rule “first adopted in the law in 1926.” His statement describes the rule that shaped the decade’s newest fortunes. When Fortune published its ranking of the seventy-six richest Americans in 1957, the top of the list was not a Rockefeller or a Mellon but an oilman, J. Paul Getty, whom the magazine credited with “a net worth of $700 million to $1 billion.” Later summaries of the list place the Texas oilman H. L. Hunt in its next tier, at $400 million to $700 million.

Dying. The estate tax was steep on paper. The IRS’s history of the tax, The Estate Tax: Ninety Years and Counting, shows a $60,000 exemption and a top rate of 77% on estates above $10 million, unchanged from 1942 to 1976. But the same 1948 legislation that created joint-return income splitting also created a marital deduction worth up to half of the adjusted gross estate, which could pass to a surviving spouse untaxed. Trusts that skipped a generation were not yet hit by any separate tax. The tax collected real money — Robert Lampman cites 36,699 estate tax returns for 1953 — but a well-advised family had trust structures and years to plan around it.

None of this was hidden. Congress set a punishing rate on ordinary income and wrote durable exceptions for the forms of income most closely tied to existing fortunes. The tax bit hardest on a high salary, which hardly any of the 267 had.

The argument over what they really paid

Here the honest answer is that serious economists disagree, and the disagreement is about definitions more than data.

On one side are Thomas Piketty, Emmanuel Saez and Gabriel Zucman, whose distributional national accounts allocate every tax — individual income tax, corporate tax, estate tax, payroll tax — to the people who ultimately bear it. On their numbers, as Zucman told the Senate Budget Committee in March 2021, “the average tax rate of the top 0.1% highest earners culminated at 60% in the early 1950s” and “remained around 55% during President Eisenhower’s two terms.” A large part of that burden is the corporate tax. They count the 52% corporate rate as falling on the owners of capital, who were overwhelmingly at the top. Zucman’s argument is that the 91% bracket mattered precisely because it made the corporate tax unavoidable: with personal rates that high, owners could not escape the corporate layer by running a business as a partnership.

On the other side are Gerald Auten and David Splinter, government tax economists — Auten at the Treasury’s Office of Tax Analysis, Splinter at Congress’s Joint Committee on Taxation. Their point is that profits retained inside a corporation are income to the owners even though they never reach a tax return, so they belong in the denominator. Count that sheltered income, and the corporate tax paid on it, as part of pre-tax income, Splinter writes, and top average tax rates in earlier decades fall significantly. In a 2020 paper in the National Tax Journal, Splinter writes that “despite a top statutory rate of 91 percent, the top 1 percent average federal income tax rate was only 16 percent” in 1962, and attributes the low figure to “extensive sheltering of income from individual-level taxes by retaining profits inside of corporations.”

These numbers do not measure the same thing — one is the top 0.1% and all taxes, the other the top 1% and federal income tax alone — so they are not a simple contradiction. Both accounts turn on the same mechanism — profits kept inside corporations while the 91% bracket stood over them. They disagree about how to allocate that income and the corporate tax paid on it — how much of the economy’s income really belonged to the top, and who ultimately bore the corporate levy. Those are questions of measurement and economic incidence, and they are not settled.

The IRS’s 1955 table settles a narrower point: for the 263 taxable returns over a million, income tax took 53% of AGI — far below 91%, far above 16%. But it describes only people who had already realised large incomes. Those who held fortunes in municipals or retained earnings are not in the table at all.

And they were most of the rich. Lampman’s estate-tax study for the National Bureau of Economic Research estimated that in 1953 about 27,000 Americans had gross estates of more than $1 million. Against 267 income millionaires two years later, that is roughly a hundred wealth millionaires for every person reporting a million-dollar income. The 91% bracket was a tax on a kind of income that almost nobody with a fortune needed to have.

What the money bought

The 1925 piece found domestic service already in retreat. By 1955 the Census Bureau could put a number on how little it paid. In its report Income of Persons in the United States: 1955, the median 1955 income of women employed as private household workers was $610, and only 29% of them worked full-time all year. The same table shows women in other service jobs with a median of $1,246. Because so many were part-timers, $610 is not the price of a full-time maid. But it is the clearest official figure for what the household-help economy paid at mid-century, and set against a million-dollar portfolio’s $30,500 coupon it shows how cheaply the staffed house could still be run at the top.

The decade’s distinctive luxury was paid for differently. If the top rate made a dollar of salary worth nine cents to the person earning it, a dollar of company-paid dinner, club dues, travel or hunting trip was worth a full dollar, deductible to the firm and untaxed to the guest. Nobody measured how much comfort ran through the expense account, but President Kennedy thought it large enough to name in his April 1961 message to Congress: “expense account living has become a byword in the American scene,” he said, asking that entertainment “and the maintenance of entertainment facilities (such as yachts and hunting lodges) be disallowed in full.” Congress did not go that far. The Revenue Act of 1962 instead added section 274 to the tax code, which required that entertainment be “directly related to” or “associated with” the active conduct of a business, that a facility be used “primarily for the furtherance” of it, that business gifts be capped at $25 per recipient a year, and that the taxpayer substantiate it all “by adequate records.”

What people get wrong

That the rich paid 91%. Nobody paid 91% of their income. The rate applied only to ordinary taxable income above $200,000 single or $400,000 joint, the whole computation was capped at 87% of taxable income, and long-term capital gains were capped at 25%. The 263 taxable returns over $1 million paid about 53% of AGI.

That nobody paid anything. Those 263 returns sent the Treasury $291 million, about $1.1 million each, and the 31 that paid at regular rates gave up 82.9% of taxable income. The bracket was real for the people it reached.

That the 1950s settle the modern tax debate. The two leading reconstructions disagree because they measure income and allocate taxes differently, not because one of them misread the data. How much of the 52% corporate tax on retained profits fell on the owners is a judgement about incidence, and reasonable economists answer it differently.

That “millionaire” in 1955 meant someone earning a million dollars. The IRS counted 267 returns over a million. Lampman estimated about 27,000 people with gross estates above a million two years earlier. The word described a fortune, and almost none of those fortunes produced a million-dollar return.

That $1 million in 1955 equals $12.5 million today. That is the price-index answer. Measured against wages it is about $21 million, and against the size of the economy about $73 million. Each is correct for a different question.

Bottom line

A million dollars in 1955 bought about $30,500 a year in top-grade corporate bonds. That is about nine average American working lives, and federal income tax took between 28% and 41% of it depending on whether the owner was married — or nothing at all, if the million sat in municipal bonds at 2.48%.

At the very top, the 91% bracket did real work. It took more than half of what 263 million-dollar earners reported, and 232 of them were already using the capital-gains route around it. What it did not do was reach the much larger group of Americans who were millionaires by wealth rather than by income, because the tax code had been written, deliberately, with room for them to stay below the line.

The lesson of 1955 is not that high rates work, or that they don’t. It is that a rate is only as broad as the income it is allowed to touch. Congress in the 1950s chose a punishing number and a narrow base, and the gap between them is where the decade’s fortunes lived.

---

Methods and sources. Return counts, AGI, taxable income, tax and effective rates are from Statistics of Income for 1955 (IRS Publication 79, 1958), Tables 1, 3, 8 and 11, read from the scanned report, checked digit by digit against its page images and reconciled with the report’s own totals (Table 8’s two tax-method rows sum exactly to the class total). Table 8’s “effective tax rate” divides tax by taxable income; the 52.8% (tax over AGI) and about 37% (36.6%: tax over AGI of the 263 taxable returns plus the excluded half of the $489.2 million net long-term gain shown in Table 11) are my arithmetic. Statutory text of the 87% limit is section 1(c) of the Internal Revenue Code of 1954 as enacted (Statutes at Large, vol. 68A). Rate schedules, standard deduction and exemptions are from the 1955 Form 1040 instructions; the bond-income tax bills are my computations. Yields are 1955 twelve-month averages of FRED AAA and the NBER Bond Buyer municipal series. Wages are the SSA average wage index; the Census Bureau’s 1955 report gives a lower average wage or salary of about $2,700, which would put the multiple nearer 11 than 9. Conversions use CPI-U (1955 average 26.775; August 2026, 334.980), the wage index (1955 vs. 2024, the latest published) and nominal GDP (1955 vs. 2025). The 1900 and 1925 multiples used railroad bond yields rather than Moody’s Aaa, so the series is not perfectly continuous. Lampman’s 27,000 is an estimate of gross estates, not net worth, built from estate-tax returns by the estate-multiplier method; the 36,699 returns are from his introduction. The two effective-rate estimates measure different groups and different taxes, as the text says. The Getty range is Fortune’s own retrospective account (Robert Lubar, “The Odd Mr. Getty,” 1986); the original 1957 issue was not consulted, The count of seventy-six names (fortunes of $75 million or more) is from Fortune’s own later account, “America’s Centimillionaires” (1968); Hunt’s $400–700 million tier comes from secondary summaries of the list. Household-worker incomes are Census P-60 No. 23, Table D. This draft was fact-checked line by line before publication.

Related reading: Time: Millionaires of 1925 · Time: Millionaires of 1900 · Taxes: How Wealth Is Structured and Preserved · The Strangest Loopholes: The Oddest Legal Ways Millionaires Pay Less Tax · Taxing the Billionaires: Wealth-Tax Debates From California to Warren · Old Money and New Money: Different Styles of Wealth