Time: Millionaires of 1900
The Million Dollar Question: In 1900, a million dollars invested in high-grade bonds produced an annual income equal to roughly how many times the average American worker’s yearly earnings?
A) About 8 B) About 25 C) About 76 D) About 200Read on for the answer.
The word has not changed in a hundred and twenty-six years. The thing it describes has changed almost beyond recognition.
“Millionaire” was already an old word in 1900 — it arrives in English from the French early in the nineteenth century — but at the turn of the twentieth it still did real work. It named a specific and very small class of people, and everybody understood roughly what it implied about how they lived. Today the same word covers close to nine per cent of American adults, describes a household that may well be house-rich and cash-poor, and implies nothing at all about staff, or travel, or how anyone spends a Tuesday.
That drift is usually explained with a single sentence about inflation, and the sentence is not wrong so much as insufficient. Prices are the smallest part of what moved. What actually changed is the relationship between a stock of capital and everything a capital owner might want to buy with it — chiefly other people’s time, which got expensive, and manufactured goods, which got cheap — and the fact that in 1900 the federal government took none of the income and almost none of the estate.
This piece is about that arithmetic, taken from the sources that were keeping score at the time.
What the word was measuring
Start with the conversion, because it is the part everyone reaches for and the part that misleads most.
Run $1,000,000 in 1900 through MeasuringWorth’s comparator and you get wildly different answers depending on what you ask. Against consumer prices, it is about $39.6 million in 2025 money — the annual CPI series runs from 8.14 in 1900 to 321.94 in 2025, a factor of 39.5, and the arithmetic is exactly that dull. Against average worker compensation, it is about $278 million. Against nominal GDP per capita — the “how rich were you compared with your neighbours” measure — it is about $323 million. Against total national output, about $1.45 billion.
Those are not four estimates of one number. They are four different questions, and the spread between them is the entire subject. A 1900 millionaire could buy about forty million dollars’ worth of goods in today’s terms and occupy a social position that would today cost hundreds of millions. Goods were the cheap part of being rich even then, relative to what came next.
So set the index numbers aside and ask what the money did.
The income, and the multiple
A million dollars in 1900 was not a spending balance. It was a claim on an income stream, and the income stream is documented.
Frederick Macaulay’s yield series for the National Bureau of Economic Research is the standard long-run record. His adjusted index of high-grade American railroad bond yields moves in a band between 3.150 and 3.202 across the twelve months of 1900, for a mean of 3.18% — the monthly detail is in his Table 10; the Census Bureau reprints only the annual low and high in Historical Statistics of the United States, 1789–1945 as series N 201–202. High-grade railroad issues are the conventional benchmark for conservative capital in this period: liquid, listed, and boring.
Government paper paid less, not more. Homer and Sylla’s A History of Interest Rates puts the longest-dated federal issue — the consol 2s of 1930 — at 1.82% in 1900, with the 4s of 1925 at 2.12%, both artificially depressed because national banks needed them as note-issue collateral. Savings banks were in between: the Comptroller of the Currency’s 1900 annual report observed that 4% had become “the exception” and that “the average rate lies between 3 and 3½ per cent.”
Call it 3.18%. A million dollars in high-grade bonds produced about $31,800 a year.
Now the denominator. Stanley Lebergott’s estimates, printed by the Census Bureau in Historical Statistics of the United States, Colonial Times to 1970 as series D 723–724, put average annual money earnings for all American employees in 1900 at $418 when employed, or $375 after deducting for unemployment. The Twelfth Census reported total manufacturing wages of $2.32 billion across 5,306,143 wage-earners for the business year 1899, which divides out to $437 — close enough to confirm the order of magnitude from an entirely separate source.
$31,800 divided by $418 is 76. The answer to the Million Dollar Question is C.
That is the whole of it. A millionaire in 1900 could live, without touching principal and without working, on an income equal to seventy-six average working lives per year. Measured against earnings net of unemployment, closer to eighty-five.
Run the same calculation now. Ten-year Treasuries have been trading around 4.68% this month, so a million dollars in the safest available paper yields about $46,800 — more, in nominal terms, than the 1900 figure. Average weekly earnings for private-sector employees ran about $1,290 in July 2026, per the Bureau of Labor Statistics, which annualises to roughly $67,100. Median household income was $83,730 in 2024.
$46,800 divided by $67,100 is 0.70. The income from a million dollars today is about seven-tenths of one average worker’s pay. It does not cover a median household.
Seventy-six, to seven-tenths. That is a factor of roughly 109, and it is a better description of what happened to the word than any inflation index.
Turn it around and it gets concrete. To generate seventy-six average incomes today you need about $5.1 million a year. At the 10-year Treasury’s 4.68% that implies roughly $109 million of capital — but that version of the sum flatters the modern investor, and the reason matters. 1900 was a gold-standard year in a broadly flat price regime, so a 3.18% coupon was close to 3.18% in real terms. A 4.68% coupon in 2026, against inflation running near 3%, is something like 1.7 to 2.2% real. Redo the calculation on real yields — $5.1 million a year at 2% — and the requirement is about $255 million, which sits very close to MeasuringWorth’s wage-based figure of $278 million.
Both numbers are defensible; they answer slightly different questions. The nominal version asks what you would have to own today to draw the same relative income. The real version asks what you would have to own to draw it indefinitely without the principal eroding, which is nearer to what a rentier in 1900 could take for granted. The second is the harder test, and the 1900 millionaire passed it.
Nothing came off the top
Here is the part that is genuinely difficult to hold in mind: that $31,800 was not taxed by the federal government at all.
Congress had tried. The Wilson–Gorman Tariff of 1894 imposed a 2% tax on incomes above $4,000. The Supreme Court destroyed it. In Pollock v. Farmers’ Loan & Trust Co., decided 8 April 1895 and reheard and expanded on 20 May, the Court held that a tax on the income of real estate or of personal property was a direct tax, that direct taxes had to be apportioned among the states by population, that this one was not, and that because the income-tax sections formed “one entire scheme of taxation,” all of them fell together.
The consequence is visible in the ledger. The Congressional Research Service’s reconstruction of federal revenue back to 1790 shows the income-tax column at zero for every year from 1873 to 1909. In fiscal 1900 the federal government took in about $567 million, of which $233.2 million was customs duties and $295.3 million was internal revenue — excise, overwhelmingly on whiskey, tobacco and beer. The order of those two is worth pausing on, because it is easy to assume the tariff was always the larger. It was not: internal revenue overtook customs in fiscal 1894 and again for the whole run from 1898 to 1902, on the back of the war taxes described below, before the tariff regained the lead in 1903. A working man’s drink and smoke funded the Republic. The bondholder’s coupon did not.
That would not change until the Sixteenth Amendment was ratified on 3 February 1913 and the Revenue Act of that October imposed 1% above a $3,000 exemption, with surtaxes rising to 6% — a combined top marginal rate of 7% on income over $500,000. Even that was gentle. In 1900 it was nothing.
State and local government was a different matter, and this is where the picture gets murkier rather than cleaner. Property taxes were the main instrument, and they nominally reached personal property including intangibles — but assessment of stocks and bonds depended largely on self-declaration, and contemporary tax commissions complained regularly that it was evaded. New York had had a state inheritance tax since 1885. What a wealthy household actually paid locally varied enormously by state and by the diligence of the assessor, and no single number describes it.
Except that something did
The neat version of this story — no income tax, no death duty, nothing — is wrong in one specific and well-documented way, and 1900 is exactly the year it was wrong. It is also wrong in a second way that almost every retelling misses, and the second correction is the interesting one.
To pay for the war with Spain, the War Revenue Act of 13 June 1898 imposed a federal legacy tax — the first federal death duty in force since 1870. It reached only personal property, never real estate. It applied only where the whole personal estate exceeded $10,000. And it graduated in two dimensions at once: by amount, and by how closely related the beneficiary was. The IRS’s own history of federal taxation of inheritance and wealth transfers sets out the grid. A child, parent or sibling paid 0.75% at the bottom; a stranger paid 5%. The statute then multiplied those base rates — by one and a half above $25,000, by two above $100,000, by two and a half above $500,000, and by three above $1 million.
Read the schedule quickly and you get a 2.25% duty on a million-dollar estate passing to a child, or 15% to a stranger. That reading is exactly what the Supreme Court threw out.
On 14 May 1900, in Knowlton v. Moore, the Court sustained the tax against the constitutional challenge — holding that it was not a direct tax at all but, in language it borrowed from an 1875 decision, “plainly an excise tax or duty” on the privilege of transmitting property rather than on the property itself. But having upheld the statute, the Court then reversed the judgment below and cut the tax’s reach back sharply. It held that the duty falls on each legacy or distributive share, not on the estate as a whole; that the progressive multiplier is therefore set by the size of each individual bequest; and that legacies not exceeding $10,000 are not taxed at all. The government had been assessing the rate off the aggregate estate. The Court told it to stop, and ordered refunds.
That distinction is worth doing the arithmetic on, because it changes the answer by an order of magnitude. Take a $1 million personal estate. The statute exempts everything passing to a surviving spouse outright, so give the widow half. Divide the remaining $500,000 among three children — about $167,000 each, which lands in the “over $100,000” band, so 0.75% doubled, or 1.5%. Each child owes about $2,500. The whole estate pays roughly $7,500, or 0.75% — not $22,500, and not 2.25%.
The tax did not last either way. Congress repealed it by the Act of 12 April 1902, effective that July, and across its whole life it raised about $14.1 million. In fiscal 1900 it produced $2.9 million — one per cent of federal internal revenue. The reasoning in Knowlton, however, outlived the statute by a century: it is the doctrinal foundation the modern federal estate tax has rested on since 1916.
Compare the modern position. The 2026 basic exclusion is $15 million per decedent, per the IRS, with a top rate of 40% above it, and the top marginal income tax rate is 37%. A millionaire in 1900 faced a federal death duty of well under one per cent in practice, nothing at all on the marital share, and no income tax whatsoever. A millionaire in 2026 faces income tax on the coupon and no federal estate tax at all, because $1 million is nowhere near $15 million.
Both centuries let the millionaire off. They do it at opposite ends.
What the money actually bought
The clearest way to feel the difference is to look at what a household of that period spent its income on, which was overwhelmingly people.
The scale of the domestic-service economy is hard to overstate. George Stigler’s NBER monograph Domestic Servants in the United States, 1900–1940 counts 1,509,000 people attached to domestic service in 1900 — 240,000 cooks, 298,000 launderers and laundresses, 104,000 untrained nurses, and 867,000 classed simply as other servants — with the caveat that the census total “is consistently low, perhaps by 10 or 20 percent.” That works out to 94.3 servants for every 1,000 private families. It is a ratio, not a share: because employing households often kept several, the proportion of families with any servant at all was meaningfully lower than 9.4%, and Stigler does not attempt that conversion. Across those four categories, 30.5% of all women in the labour force were doing this work.
And it was cheap. The one comprehensive wage study of the period is Gail Laughlin’s survey for the United States Industrial Commission, published in 1901, which drew returns from thirty-nine states and three territories. Her own national averages for general servants are $3.28 a week from employers’ returns and $3.33 from employees’. Stigler recomputed them from the thirty-three states with usable data, weighting by servant counts, and got $3.16 a week — about $164 a year, plus room and board. He notes that the collection date is not stated anywhere in the report and that 1899 is a guess. Whichever figure you take, it varied sharply by region: $4.13 in the West, $3.34 in the Northeast, $2.22 in the South. Stigler is also blunt that the early wage evidence is thin, writing of exactly this material that it may be representative “and this is questionable.”
Lucy Salmon’s Domestic Service put national averages at $7.18 a week for men and $3.23 for women, with butlers at $6.11 in her employer schedules — but those returns were collected in 1889 and 1890, so they describe the decade before, not 1900, and the 1901 reprint makes them look more current than they are. Hours, per Laughlin, split revealingly by who was answering: employers reported about twelve hours on call and nine at actual work; servants reported thirteen on call and twelve working.
Great houses paid several times the going rate, and there is a source for that too. Mary Elizabeth Carter’s Millionaire Households and Their Domestic Economy, published in 1903, is a management manual for exactly this readership, and it prints wages — in words rather than figures, which is why they are rarely quoted. A superintending housekeeper commands “from fifty to one hundred and fifty dollars a month.” A chef, “about one hundred dollars a month,” plus commissions on the market bills that Carter treats as an unremarkable fact of the trade. A lady’s-maid, twenty-five to forty. A parlor-maid, twenty to thirty. A housemaid, eighteen to twenty-five. A superior laundress “seldom receives over thirty dollars a month,” her assistants down to eighteen. A dining-hall maid “seldom gets more than sixteen dollars a month.”
Carter never gives a butler’s salary, or a footman’s. For the butler she gives a ratio instead: his wages “equal hers” — the parlor-maid’s — “multiplied three or four times,” which puts him somewhere around $60 to $120 a month. She is firm about headcount. “In order to establish an unmistakable air of quiet ease in a millionaire’s mansion the butler should have at least four men with him.” For a dinner of twenty-four, “eight men, and the butler as major domo, making the ninth.” The servants’ hall she specifies should seat at least twenty-four at once, sized to the establishment.
Build a payroll from her figures and a twenty-person indoor staff comes to roughly $845 a month, about $10,100 a year: a housekeeper at $150, a chef at $100, a butler at $90, four footmen at $45, a lady’s-maid at $35, two parlor-maids at $25, four housemaids at $22, a laundress at $30 with two assistants at $18, a dining-hall maid at $16, two useful men at $35. Two of those lines are mine rather than hers — the butler at $90 is the midpoint of her three-to-four-times ratio, and the footmen at $45 have no basis in the book at all, which matters because they are the second-largest item in the table. Wages only; add board, uniforms and the outdoor staff and it climbs.
Against $31,800 of bond income, twenty resident servants absorb about a third of the year’s money and leave two-thirds for the house, the horses, the travel and everything else.
Now the same household in 2026. Put twenty full-time domestic employees at a fully loaded $85,000 each — illustrative rather than surveyed, and low for live-in staff in a wealthy household — and you are at $1.7 million a year, thirty-six times the entire annual income of a million-dollar portfolio. Put the other way: $46,800 a year buys about half of one household employee.
Servants and space became dearer; goods and travel became cheaper. That divergence is not a footnote to the story of the 1900 millionaire. It is the story.
Nobody counted them
One more thing about 1900 that is easy to miss: there is no reliable count of how many millionaires there were, and there could not have been.
There was no income tax, so there were no returns. There was no estate tax before June 1898 and none after mid-1902, and the one in between reached personal property only. There was no wealth register, no Forbes list, no bank reporting regime. Anyone who tells you the number is quoting an estimate built from newspapers and gossip.
The best of those estimates is itself a remarkable document. Beginning in October 1890, the New-York Tribune spent more than a year and a half canvassing merchants, bankers, commercial agencies, lawyers and county surrogates, and in 1892 published American Millionaires: The Tribune’s List of Persons Reputed to Be Worth a Million or More — 4,047 names, arranged with the line of business in which each fortune was made. Note the word on the title page: reputed. Note also why the paper did it. The introduction says the work was undertaken in the public interest “especially of the Republican party of the United States,” as a direct rebuttal to a Farmers’ Alliance claim that there were 31,000 American millionaires, in the middle of a tariff fight. The threshold is elastic — it takes in people “in very close proximity to a million.” Treat 4,047 as the Tribune’s claim, not as a count.
The commonly repeated figure of four to five thousand millionaires in 1900 is essentially that 1892 compilation, aged eight years by assumption. Against roughly 16.2 million American families in the 1900 census, four thousand names is on the order of one per four thousand families, or 0.025% — bearing in mind that the Tribune counted individuals and some undivided estates, not households.
The contrast with now is not only in the number but in the fact that we have one. UBS’s Global Wealth Report put the United States at 23.6 million dollar millionaires at the end of 2025, counted as adults. Against a US adult population of about 270 million, that is 8.8% — roughly one adult in eleven and a half, rather than 0.025% of families. The methodology is contestable and the definition includes home equity and retirement accounts, but the estimate exists, is produced annually, and rests on national balance-sheet data that simply did not exist in 1900.
What people get wrong
That inflation explains it. Prices multiplied by about 39.5 between 1900 and 2025. The gap between a 1900 millionaire’s economic position and a 2026 millionaire’s is roughly 109x on an income-to-average-wage basis. Those combine multiplicatively, not additively: 109 is 39.5 times 2.76. Inflation is the larger of the two factors, but the second one — the average American getting substantially richer — is doing about a third of the work, and it is the one nobody mentions.
That there was no federal tax on wealth in 1900. There was, and 1900 is precisely the window in which it operated. The 1898 legacy tax ran from June 1898 to July 1902 on personal property only. That it was small, temporary and, after Knowlton, assessed legacy by legacy does not make it imaginary — it was the first federal death duty in force since 1870 and the direct doctrinal ancestor of the one in force today.
That it therefore hit big estates hard. It did not. Bequests to a surviving spouse were exempt outright, legacies of $10,000 or less were exempt after Knowlton, real estate was never touched, and the progressive multiplier attached to each individual share rather than to the estate. A million-dollar personal estate left to a widow and three children paid on the order of $7,500. The headline 2.25% and 15% figures describe the schedule, not the outcome.
That the 1894 income tax was repealed. It was struck down, which is a different thing with different consequences. Pollock did not merely end one statute; it held that taxing income from property was a direct tax requiring apportionment, and that holding required a constitutional amendment to undo. Eighteen years passed between the decision and the Sixteenth Amendment.
That a millionaire’s money sat in stocks. For conservative capital in 1900 the benchmark was bonds — high-grade railroad issues at 3.18%, governments below 2%, savings deposits at 3 to 3.5%. Equity ownership of the sort that dominates modern portfolios was a promoter’s business, not a widow’s. The word “rentier” was not yet an insult.
That servants were a luxury. They were the way domestic work got done, at a rate of 94 per thousand families and a mean wage of about $3.16 a week. What distinguished a millionaire’s house was not that it had a servant but that it had twenty, at three or four times the going rate, arranged in a hierarchy that Carter’s book exists to explain. The luxury was the org chart.
That the comparison can be made cleanly at all. It cannot, quite. The 1900 millionaire had no antibiotics, no air conditioning, no aeroplane, no reliable refrigeration outside an ice-box, and — smallpox aside — no vaccine a civilian would routinely receive. Life expectancy at birth was about 47 years, though that figure is driven mostly by infant mortality and therefore describes a wealthy adult’s prospects badly. Any statement that $1 million in 1900 “equals” some figure now is a statement about one dimension chosen from several, and the honest version names the dimension.
Bottom line
A million dollars in 1900 bought an income of about $31,800 a year, untaxed by the federal government, roughly seventy-six times what the average American worker earned, in a country where a resident servant cost about $164 a year plus board and there were perhaps four thousand people in the category.
A million dollars in 2026 buys a taxed income of about $46,800, roughly seven-tenths of what the average American worker earns, in a country where a single household employee costs more than twice that and there are 23.6 million people in the category.
The nominal income went up. Everything the income is measured against went up far more.
What makes 1900 worth looking at is not that the rich were richer — by consumption they were poorer than a comfortable household today, and they died younger. It is that the distance was larger, and that the federal government took essentially none of it. That second condition was already on a clock. The legacy tax expired two years later, the Sixteenth Amendment arrived thirteen years later, and the top marginal income tax rate reached 77% by 1918. The 1900 millionaire was standing at the end of a long era and had no way of knowing it.
The lesson generalises past the period. A wealth threshold is a moving object with three variables underneath it — the price level, the yield on capital, and what everyone else earns — and only the first of them is what people mean by inflation. When someone tells you a number is what it takes to be rich, ask which of the three they have held still.
Methods and sources. Bond yields are Frederick Macaulay’s adjusted index of high-grade American railroad bond yields (NBER, 1938); the 3.18% figure is the twelve-month mean of his monthly Table 10, not a published annual average — Census series N 201–202 in Historical Statistics of the United States, 1789–1945 reprints only the 1900 low (3.150) and high (3.202). Government and savings-bank rates are from Homer and Sylla’s A History of Interest Rates (4th ed.) and the 1900 Comptroller of the Currency report. Earnings are Stanley Lebergott’s estimates as printed in Historical Statistics of the United States, Colonial Times to 1970, series D 723–724; the $437 manufacturing figure is derived by division from Twelfth Census totals for the business year 1899, not printed by the census. Servant counts and the 30.5% female-labour-force share are from Stigler’s NBER monograph; the $3.16 weekly wage is Stigler’s own reweighting of Gail Laughlin’s 1901 report to the Industrial Commission, not Laughlin’s published average, which was $3.28–$3.33. The 94.3-per-1,000 figure is a ratio of servants to private families and is not a share of families employing one; no source supports that conversion and this piece does not make it. Salmon’s wage figures were collected in 1889–90 and are labelled as such. Household wages are quoted verbatim from Mary Elizabeth Carter’s Millionaire Households and Their Domestic Economy (1903); the twenty-person payroll built from them is my construction, the butler’s $60–120 range is inferred from her three-to-four-times ratio, the $45 footman rate has no basis in the book, and Carter gives no total household budget anywhere — do not attribute one to her. The $85,000 modern staff cost is illustrative, not surveyed. Tax figures are from Pollock, Knowlton v. Moore, the IRS history of inheritance taxation and CRS report RL33665; “plainly an excise tax or duty” originates in Scholey v. Rew (1875) and is quoted by the Knowlton Court. The FY1900 receipts split of $233.2 million customs to $295.3 million internal revenue is confirmed in two independent federal sources. The count of millionaires rests on the New-York Tribune’s 1892 compilation of 4,047 reputed millionaires, assembled for an avowedly partisan purpose and counting individuals rather than households; no count exists for 1900 itself, and none could, because no federal instrument collected the data. Inflation conversions are MeasuringWorth, 1900 to 2025. This draft was fact-checked line by line before publication and the check changed the argument in one central place: an earlier version read the 1898 legacy tax schedule off the whole estate, producing a 2.25% duty on a million-dollar bequest to a child. Knowlton v. Moore — decided in May 1900, and cited in the same section — held precisely the opposite, that the rate attaches to each legacy and that the government’s aggregate method was unlawful. The corrected figure is roughly a tenth of the original, and the exemption for a surviving spouse, missing entirely from the first draft, matters more than the rate schedule does. Fifteen further figures and characterisations were corrected in the same pass.
Related reading: Wealth Levels: Life at $1M, $10M, $100M, and $1B · Taxes: How Wealth Is Structured and Preserved · Staff: Outsourcing Daily Life · Old Money and New Money: Different Styles of Wealth · The Job That Sees It: The Estate Manager
