In its early days, the federal income tax was something of a mystery. Policymakers expected the levy to raise money and reallocate the tax burden. But for the first couple of years, that was just a guess; members of Congress were lawmaking by the seat of their pants.
From 1913 until 1916, Congress lacked answers to even the most basic questions. Who was actually paying the individual income tax? How much were they earning? From what sorts of jobs? And most important, at least from the perspective of elected officials: Where did these taxpayers live?
The Bureau of Internal Revenue (BIR) was slow to answer these questions. But in 1918 the agency began to sketch a portrait of the American taxpayer: Judged by probabilities, a person paying the nascent income tax was disproportionately likely to be (1) rich; (2) working as a merchant; or (3) living in New York.
Congressional Mandate
After several years of guesswork, Congress directed the BIR to get serious about data collection. “To the Congress, framing the Revenue Act of 1916, it became apparent that accurate information regarding the distribution of income in the United States was necessary,” explained Edward White, chief of the BIR Statistical Division between 1918 and 1946.
Section 21 of the 1916 revenue law required the BIR to publish an annual report featuring “statistics reasonably available with respect to the operation of the income tax law.” More specifically, lawmakers asked for data “classifications of taxpayers and of income, the amounts allowed as deductions and exemptions, and any other facts deemed pertinent and valuable.”
The BIR’s first annual report appeared two years later, when the agency published the inaugural edition of “Statistics of Income.” The report, which drew most of its data from returns filed in 1916, focused on the amount of taxes paid, the amount of net income that produced this tax revenue, and the location where returns were filed.
“In fact, a major portion of each book related to geography,” White explained. Ever sensitive to the needs (and whims) of their constituents, politicians were eager to discover which states were paying the most. That preoccupation was hardly surprising, given the long-standing importance of regional arguments about income taxation. Generally speaking, lawmakers from the agricultural south and west had been supporting the levy as a means to shift more of the tax burden to the industrial northeast.
Subsequent BIR statistical reports grew in both scope and size. In 1918 Congress expanded the reporting requirement to include data on special wartime levies, including the excess profits tax. But more broadly, the reports began offering general economic data in addition to facts about tax collection specifically. As White explained:
Such development was supported and encouraged because (1) the cost of preparing the data from documents already filed for administrative purposes was relatively low, (2) more stringent filing requirements continually produced a better statistical universe of individual returns, and (3) as the Internal Revenue Code became more complicated, the return form itself contained more desirable data.
For the first decade or so, the bulletin derived its data strictly from returns showing taxable income; comprehensive data culled from all returns, including those with no tax liability, did not appear until 1928.
The development of BIR statistical reports proved important, at least to a certain kind of early-20th-century stats wonk. “These statistical reports have inaugurated an epoch in income statistics,” White wrote in a 1928 essay for The Accounting Review:
Never before had the economist, the statistician, or the business executive either a contemporaneous or historical presentation of the financial status of the civil organization of a nation so interesting, so valuable, or so helpful in determining the distribution of incomes, the rise and fall of profits, or the purchasing power of communities.
When it came to data, White insisted, the BIR had a responsibility to remain objective. “Its functions are not to indulge in ethical speculation, abstract theories, or personal philosophies,” he insisted during a 1921 speech to the Monday Lunch Club of Washington, D.C. “Its economic service is that of providing Congress with concrete information as to classes of taxpayers, sources of income or nature of business pursuits; or preparing special compilations for use in the administration of the Income Tax Laws.”
White also contended that the BIR owed something to the public. Through its publications, he said, the agency was providing “public information of value, not only in the study of the economic conditions of the country, but likewise facts and figures upon which and through which the business statistician and business administrator may chart comparisons of productions and gauge the potential absorption power by geographical divisions.”
Year 1
White was correct: Accurate data about the federal tax system was much in demand by the time the BIR issued its first “Statistics of Income” report. Newspapers around the country were quick to scour the publication for information relevant to their (typically well-heeled) readers.
World War I had played a key role in boosting popular interest in tax data. When the United States entered the war, Congress responded by lowering exemptions and raising rates for the personal income tax (among many other revenue-increasing tax reforms). After two years, Americans were eager to know how such changes were playing out.
“What sort of people and what sort of industries will pay the principal cost of running the war?” asked The Brooklyn Daily Eagle in 1918. And the BIR had finally delivered an answer — albeit an imperfect one.
Since it was based on 1916 data, the 1918 volume of “Statistics of Income” couldn’t directly illuminate the distribution of the wartime tax burden; the United States didn’t enter the fighting until the following year, and tax legislation was always playing catch-up relative to spending needs.
But the 1916 data did provide clues about the ways in which higher taxes might be distributed across regions, professions, industries, and income classes. “People who paid certain amounts in 1916 will pay next year a great deal more but the relative paying strength of their different occupations will not show much change,” The Brooklyn Daily Eagle contended. “The higher taxes are falling upon people in all walks of life, just as they did in 1916, and with the same general impartiality.”
Of course, income taxes did not fall upon people “in all walks of life.” In 1916 the income tax was still a rich person’s problem — just as lawmakers had intended when they imposed the levy three years earlier. Congress had established notably high exemptions: $3,000 for individuals and $4,000 for married couples ($85,900 and $115,000 in 2024 dollars, using the inflation adjustments at MeasuringWorth.com).
For tax year 1917, however, Congress had opted to cast a wider net, lowering exemptions for both single and married filers to $1,000 and $2,000, respectively ($23,800 and $47,500 in 2024 dollars). This change did not make the income levy a true “mass tax”; such a transformation would take another 25 years and the arrival of a second world war.
Still, the exemption cuts for tax year 1917 dramatically expanded the number of people paying income taxes. The number of taxable individual returns increased nearly eightfold, rising from 437,036 in 1916 to 3,472,890 in 1917.
In other words, the income tax of 1917 was very different from the income tax of 1916. But given the paucity of available statistical data, it still seems reasonable that newspaper editors were poring over the old numbers looking for clues about the new ones.
A Picture in Numbers
So what did the BIR ultimately tell Congress (and the nation) about those 437,036 returns filed in 1916?
In aggregate, taxpayers reported net income of $6,298,577,620. Their combined tax payments totaled $173,386,694, yielding an average tax of $396.73. The average tax rate on this net income was 2.75 percent.
As expected, people paying the income tax were generally well-off — and many were downright rich. To again use the same framing as the BIR report:
26.4 percent of total net income was reported by filers earning between $3,000 and $10,000 ($85,900 to $286,000 in 2024 dollars).
49.5 percent of net income was reported by people earning between $3,000 and $30,000 ($85,900 to $859,000 in 2024 dollars).
76.2 percent of total income was reported by taxpayers earning less than $150,000 ($4.3 million in 2024 dollars).
In terms of the number of returns filed, 98.5 percent came from people earning less than $100,000 annually, the BIR observed. But those earning more than $100,000 — people the agency described as “the 1 1/2 percent” — reported nearly one-third of total net income and nearly three-fourths of the total tax paid.
Many types of demographic data were hard to distill from tax returns. In particular, the BIR warned, “it is not possible to classify taxpayers accurately by sex and conjugal conditions, owing to the fact that 81.25 per cent of the returns are made by heads of families including joint returns of husband and wife.”
Still, the BIR could shed a bit of light on this subject, noting that 7,635 returns came from married women filing separately (1.75 percent of all returns filed). Single male taxpayers filed 47,461 returns (10.86 percent of the total), and single women filed 26,833 returns (6.14 percent of the total).
Occupations
The BIR offered a partial breakdown of returns according to occupation. The selection was a bit idiosyncratic, focusing especially on jobs “in which a relatively small number of minors are employed.” But in any case, the largest number came from “Merchants and dealers,” who filed 54,363 returns. After the merchants came “Manufacturers” (23,631 returns), “Lawyers and judges” (21,273 returns), and people in the “Medical profession” (20,348 returns).
Some professions that didn’t account for an especially large number of total returns still filed a lot of returns relative to the number of people in the category. Fully 20.68 percent of people working as “Stock and bond brokers” filed returns in 1916, as did 20.77 percent of “Brokers — All other.” By contrast, all those returns coming from “Merchants and dealers” represented just 4.36 percent of those working under that designation.
The BIR tried to assess the relative importance — for tax purposes — of incomes derived from personal service, business, and property. The divisions were a bit arbitrary but still illuminating. “Under personal service have been classified incomes and salaries, professions and vocations,” the agency explained. “Partnership gains and profits of all kinds have been included in the class devoted to business pursuits.” Finally, the BIR classified “rents, royalties, interest income reported by fiduciaries, income from foreign sources, and dividends” as “income from property.” (The agency noted that this set of calculations was based on gross income, since it was not possible to allocate deductions according to this scheme.)
More than half of the gross income had come from a combination of personal services (22.17 percent) and business profits (31.59 percent). Dividends also made a large contribution: 25.59 percent of total gross income reported in 1916 returns.
In trying to assess the relative importance of income from labor and property, the agency also came to a striking if unsurprising conclusion: “It will be noted that the proportion of property income steadily rises as the income itself becomes larger,” the BIR wrote.
Geography
Finally, the BIR offered a geographical breakdown of personal and corporate income taxes, combining both to offer a single ranking. Fully 35.47 percent of the total tax paid nationally had come from New York — a fact that did not escape the attention of the editors at The Brooklyn Daily Eagle. Extrapolating from the 1916 data to estimate the likely burden of pending legislation, the paper declared, “New York Will Pay $825,000,000 of the New Revenue Bill.”
Other states paying heavily in 1916 could also expect a heavy burden after Congress finished work on the new revenue bill. These included Pennsylvania (11.55 percent of total tax paid in 1916), Illinois (7.52 percent), Ohio (5.82 percent), and Massachusetts (5.74 percent).
Near the bottom of the list were North Dakota, having paid just 0.079 percent of the total national tax bill, Wyoming (0.069 percent), and South Dakota (0.056 percent). Bringing up the rear was Nevada, footing just 0.027 percent of the total.