The four questions that sort every tax dispute

Few fields of American law produce as many confident falsehoods as the revenue code. The statutes run long, the forms intimidate, and the money involved belongs to the reader, so folk theories about what the law requires, whom it favors, and what it has accomplished circulate with unusual energy. Some of these theories flatter the skeptic, some flatter the defender of the existing system, and the honest response to both is the same: check the statute, check the decisions, check the published numbers, and report what they support.

This article applies one test to a dozen of the most repeated contentions in American tax debate. A reader arrives holding one of the claims that circulate about the national levy on earnings, on wages, and on business, and leaves knowing what the statutes, the courts, and the published data actually support. Each verdict is sourced. Claims get corrected in every direction at equal length, because a correction that only runs one way is advocacy, not competence. And where the evidence genuinely does not settle a question, the article says so, with both sides named, rather than forcing a verdict the record cannot carry.

Almost every disagreement in this field comes down to four ambiguities, and naming them resolves more disputes than any argument about percentages ever will. The four-question sort: almost every tax dispute reduces to which tax is being discussed, which measure of income, which unit of analysis, and which time period, and naming those four resolves more disagreements than any argument about rates. Consider how much confusion each question prevents. Which tax is being discussed matters because the individual income levy, the payroll levy, the corporate levy, and the estate levy have different bases, different rates, and different distributional footprints, and a statement that is true of one is routinely misapplied to another. Which measure of income matters because adjusted gross income, cash income, and broader economic definitions of earnings produce different pictures of who earns what. Which unit of analysis matters because individuals, tax units, and households sort the same population into different groups. Which time period matters because a single year and a lifetime tell different stories about burden. Keep those four in view and the claims below become far easier to grade.

The verdicts use four labels. True means the statute, the decisions, or the dataset support the claim as stated. Partly true means the claim holds under some stated condition and fails under others. False means the authorities contradict it. Genuinely unresolved means credible evidence and credible experts point in different directions and no honest reading can declare a winner. That last label is not a dodge. On questions of distribution and incidence, the economics profession genuinely disagrees, and pretending otherwise would repeat the exact error this series exists to correct.

The sources throughout are primary where possible. The Internal Revenue Code supplies the statutory text. Supreme Court and appellate decisions supply the judicial answers. The Congressional Budget Office, the Joint Committee on Taxation, the Treasury Department, and the Internal Revenue Service Statistics of Income division supply the published numbers. Secondary commentary appears only to explain what the primary sources mean, never to substitute for them. Everything below is dated to September 2013 or earlier, or is plainly historical, except for two explicitly dated later data points that the series date-wall rule permits, and nothing here is individualized guidance. The series thesis for this installment is simple: correcting the record is part of making a reader competent, and competence means knowing which claims the record supports, which it refutes, and which it leaves open.

Two illustrations show the four questions at work before the claims begin. Take the contention that the wealthy pay no tax. Ask which tax is being discussed: the statement might gesture at the income levy while ignoring payroll and excise levies that the same households pay. Ask which measure of earnings is in play: a definition that excludes unrealized gains will show different top shares than one that includes them. Ask which period the figures cover: a single loss year for one company or one investor proves nothing about a decade. Each answer changes the verdict, which is why the claim cannot be graded until the questions are answered. Take the contention that corporations pay no tax. Ask which companies and which years: the sector’s aggregate collections across a decade tell a different story than one company’s return in a loss year. Ask who bears the burden: the check written by the corporation is not the same as the economic incidence. The four questions do not supply answers by themselves, but they expose which answer a given piece of evidence can actually support.

The phenomenon has a name in legal scholarship: folk law, the body of confident legal belief that circulates among non-lawyers and bears only a passing resemblance to the law of the books. Folk tax law has its own greatest hits, and several appear in this article. Folk law persists because the real law is costly to learn and the folk version is free, and because the folk version usually contains a grain of truth that makes it feel verified. The voluntary myth has its grain in the Service’s own phrase. The bracket myth has its grain in the pre-indexing era’s bracket creep. The personal account myth has its grain in the annual earnings statement. Each grain is real, and each inference drawn from it is wrong, which is why debunking requires more than a verdict: it requires replacing the grain with the full mechanism, as the sections below attempt.

A word on method. Each claim below is stated first in the strongest form its holders would recognize, before any correction begins. That discipline matters because weak versions of opposing arguments are easy to defeat and teach nothing. The corrections then run in every direction at equal length: the constitutional and skepticism flavored claims receive the same space and the same care as the claims about distribution and revenue. No characterization of claim holders appears anywhere below. The verdicts rest on statutes, decisions, and named datasets, and where the evidence does not settle a question, the article says so with both sides named. The point is not to leave the reader with approved opinions about rates. The point is to leave the reader able to grade the next claim without assistance.

The frivolous position list operates as a warning system worth understanding on its own terms. The Internal Revenue Service publishes and periodically updates notices identifying arguments it considers frivolous, and section 6702 penalties apply to submissions based on listed positions. The list exists so that filers encounter the government’s answer before they file, not after. Taxpayers who disagree with the list are free to litigate the underlying merits; what they cannot do is plead ignorance of the agency’s view when penalties are assessed. Courts have uniformly sustained the scheme against challenges, treating it as fair notice rather than prejudgment.

State income taxes add a final layer of reliance. Most states that levy their own tax begin their computations from federal adjusted gross income or federal taxable income, incorporating the federal definition by reference. That piggybacking means the constitutional theories, if ever accepted, would unwind not only the federal system but the revenue structures of more than forty states. No court has been willing to take that step on the strength of punctuation discrepancies, and the state courts that have faced the theories have rejected them on the same grounds as the federal courts.

Claim one: the Sixteenth Amendment was never ratified, and wages are not income

Stated in the strongest form its holders would recognize, this contention runs as follows. The Sixteenth Amendment, which its supporters cite as the foundation of the modern income levy, was never properly ratified: the ratification proclamations were defective, some states approved text that differed from what Congress proposed, and therefore the amendment is void. Alternatively, even if the amendment stands, it does not authorize a levy on wages, because wages are an equal exchange of labor for money rather than a gain, and the word income in the amendment means only gain or profit. On either version, the conclusion is the same: the national government lacks constitutional power to tax what ordinary people earn by working.

The argument fails at every step, and it is worth explaining why with care rather than dismissal, because the confusion underneath it is understandable. Start with the constitutional background. Before 1913, Article I of the Constitution required that any direct tax be apportioned among the states by population. In Pollock v. Farmers Loan and Trust Company, 157 U.S. 429 (1895), the Supreme Court held that a tax on income from property was a direct tax subject to that apportionment rule, which in practice made a general income levy unworkable. Congress responded by proposing the Sixteenth Amendment, which provides that Congress shall have power to lay and collect taxes on incomes, from whatever source derived, without apportionment among the several states. Ratification was completed on February 3, 1913, when the thirty-sixth of the forty-eight states then in the Union approved it, and Secretary of State Philander C. Knox proclaimed the amendment adopted on February 25, 1913.

The ratification challenges center on clerical discrepancies: punctuation, capitalization, and wording differences between the text Congress proposed and the texts some state legislatures approved. Holders of this view have catalogued these differences at length and argued that they invalidate the approvals. Courts have rejected this position uniformly. The leading appellate treatment is Miller v. United States, 868 F.2d 236 (7th Cir. 1989), which examined the discrepancies and held that they did not defeat ratification, noting that the political branches had long treated the amendment as valid and that the variations were immaterial. The Tenth Circuit in United States v. Collins, 920 F.2d 619, 629 (10th Cir. 1990), called the contention that the amendment does not authorize a direct, nonapportioned income tax “devoid of any arguable basis in law,” and the Supreme Court denied review, 500 U.S. 920 (1991). The Fifth Circuit, in Crain v. Commissioner, 737 F.2d 1417, 1418 (5th Cir. 1984), a per curiam opinion, wrote that it perceived “no need to refute these arguments with somber reasoning and copious citation of precedent; to do so might suggest that these arguments have some colorable merit.” No federal court has ever held the amendment invalid on ratification grounds. The Supreme Court, for its part, treated the amendment as effective from the beginning, building an entire body of income tax jurisprudence on it starting with Brushaber v. Union Pacific Railroad Company, 240 U.S. 1 (1916), decided unanimously on January 24, 1916.

Brushaber deserves attention because it answers the second version of the claim as well. The Court held that the amendment did not create a new power to tax but removed the apportionment obstacle Pollock had erected, restoring to Congress the full scope of its taxing power over incomes. Stanton v. Baltic Mining Company, 240 U.S. 103 (1916), decided the same year, confirmed that the amendment authorized taxes on income without apportionment and did not convert the income tax into something the Constitution forbids. The distinction the amendment draws is between direct taxes requiring apportionment and taxes on incomes, which it expressly exempts from that requirement. Nothing in the amendment, and nothing in the decisions construing it, limits income to gains from property or excludes compensation for labor.

The statute is explicit on wages. Section 61(a)(1) of the Internal Revenue Code defines gross income to include compensation for services, including fees, commissions, fringe benefits, and similar items. The Supreme Court has read income broadly for the better part of a century. In Eisner v. Macomber, 252 U.S. 189 (1920), the Court defined income as gain derived from capital, from labor, or from both combined, expressly including labor. In Commissioner v. Glenshaw Glass Company, 348 U.S. 426 (1955), the Court adopted the now standard formulation: undeniable accessions to wealth, clearly realized, and over which the taxpayer has complete dominion. Wages fit that definition without strain. Earlier, in Lucas v. Earl, 281 U.S. 111 (1930), the Court held that earnings from personal services are taxed to the person who earns them, rejecting an attempt to shift them by private agreement. The proposition that wages are not income has been raised in court many times and has lost every time, with appellate decisions describing the argument as frivolous and imposing sanctions on those who press it.

Those sanctions are real and worth stating plainly. Section 6702 of the Code imposes a 5,000 dollar penalty on frivolous tax submissions, including returns that rest on positions the Internal Revenue Service has identified as frivolous, and the agency has repeatedly listed Sixteenth Amendment and wages are not income arguments in that category. The 5,000 dollar figure has applied since 2006, when Congress raised the penalty from 500 dollars; it is the correct figure for the period this article covers. Section 6673 permits the Tax Court to impose penalties when a taxpayer advances frivolous positions in litigation. Courts of appeals routinely affirm these penalties and sometimes add their own sanctions for frivolous appeals. On the criminal side, Cheek v. United States, 498 U.S. 192 (1991), held that a good faith misunderstanding of the law can negate the willfulness required for criminal tax liability, but the decision did not validate any of the underlying theories, and civil liability and penalties remain. The practical consequence is that advancing these arguments in filings or in court costs money and changes nothing about the underlying obligation.

Why does the contention persist despite this unbroken record of rejection? Part of the answer is textual. The Constitution’s apportionment language is genuinely confusing to a modern reader, and Pollock’s distinction between direct and indirect levies invites the thought that some clever categorization might exempt earnings from labor. Part of the answer is the amendment’s spare wording, which says nothing about wages specifically and therefore leaves room for motivated reading. And part of the answer is that the ratification discrepancies are real as clerical facts: states did approve texts with minor variations. The error lies in the leap from clerical variation to legal invalidity, a leap no court has accepted, because constitutional law has never required letter perfect uniformity in state ratification resolutions, and because a century of reliance has settled the question beyond reopening. Verdict: false, decided by the amendment’s text, the certification of 1913, Brushaber and its progeny, section 61, and the uniform rejection by the federal courts, with statutory penalties attaching to the argument’s use in filings.

The deeper history makes the rejection easier to understand. The federal government first taxed incomes during the Civil War, and the Supreme Court sustained that levy in Springer v. United States, 102 U.S. 586 (1881), treating it as an indirect tax not subject to apportionment. The trouble began with the Wilson-Gorman Tariff Act of 1894, which imposed a 2 percent tax on incomes above 4,000 dollars. In Pollock, a divided Court held 5 to 4 that taxes on income from property were direct taxes requiring apportionment, effectively killing the statute, since apportioning an income tax by state population is unworkable. Pollock did not hold that Congress lacked power to tax incomes; it held that one category of income taxes had to be apportioned. The Sixteenth Amendment was drafted to remove that obstacle with surgical precision, and its language tracks the problem Pollock created: taxes on incomes, from whatever source derived, without apportionment.

The ratification drive that followed was deliberate and public. President Taft proposed the amendment in 1909, Congress approved the text, and the states took it up over the next four years. Thirty-six of the forty-eight states ratified, exactly the three-fourths required, with Delaware, Wyoming, and New Mexico completing the count on February 3, 1913, and Secretary of State Knox proclaimed ratification on February 25, 1913. The clerical discrepancies that later drew attention, differences in punctuation, capitalization, and a few words between the congressional resolution and some state approvals, arose from the ordinary friction of enrolling and copying texts by hand across dozens of legislatures. A self-published volume catalogued them decades later and argued they voided the approvals. The argument misunderstands what ratification requires: the states approved the substance of the proposed amendment, the political branches accepted those approvals, and constitutional practice has never demanded letter perfect uniformity across state resolutions.

Part of the persistence comes from the genuine oddity of the direct tax clauses. The Constitution never defines direct tax precisely, the founders’ own understanding was contested, and Pollock’s 5 to 4 split shows how close the question was. Readers who encounter the apportionment language for the first time reasonably wonder whether the income tax fits it. The answer the legal system settled on is that the amendment carves income taxes out of the apportionment requirement by name, which ends the inquiry no matter how one classifies the underlying levy. The amendment does not repeal the apportionment rule for other direct taxes; it exempts taxes on incomes from it. That is why arguments built on classifying the income tax as direct or indirect miss the point: the amendment makes the classification irrelevant for income.

The doctrinal answer on wages runs through the Court’s definition of income. Eisner v. Macomber described income as gain derived from capital, from labor, or from both combined, which places compensation for work inside the definition by its own terms. Later decisions refined the formulation without narrowing it: Glenshaw Glass asks only for undeniable accessions to wealth, clearly realized, under the taxpayer’s complete dominion, a test wages meet on the day the paycheck clears. Lucas v. Earl added the assignment principle, taxing earnings to the earner regardless of private agreements to redirect them. The folk theory that labor exchanged for money is an equal trade producing no gain has been offered to courts in many forms, sometimes dressed in the language of basis and cost, and it has been rejected in every form, because the Code measures gain on labor at zero basis: the worker’s time is not a capital investment with a cost to recover.

Related variants circulate alongside the ratification theory and fail for the same reasons. One holds that the amendment was never properly proclaimed because of defects in Knox’s certification; courts treat the proclamation as conclusive on the political question of ratification. Another holds that only certain kinds of income, such as corporate profits, count; section 61 and the decisions above answer that directly. Each variant, examined on its own terms, collapses into the same two errors: a clerical discrepancy mistaken for a legal defect, or a narrow reading of income contradicted by the statute’s text.

The variant arguments form a catalog of their own, and the courts have worked through it methodically. Filers have argued that they are not persons within the meaning of the Code, that Federal Reserve notes are not dollars and therefore not income, that the Fourteenth Amendment created a separate class of citizenship exempt from direct levies, and that filing is required only of federal employees or residents of the District of Columbia. Each theory has its own small literature of pamphlets and seminars, and each has been rejected in published opinions, often with the same sanctions. The Tax Court maintains that a position is frivolous when it is contrary to established law and unsupported by a reasoned argument for changing it, a standard these theories cannot meet because the law they contradict includes Supreme Court decisions directly on point. Repeat filers have been declared vexatious litigants, barred from filing further suits without leave, and promoters have been enjoined from selling the schemes. The uniformity is the point: there is no jurisdiction, no judge and no era in which these arguments have succeeded, which is why the article states the verdicts without hedging and without ridicule.

The uniformity of the judicial rejection is itself evidence. Courts of appeals in every circuit, the Tax Court, and the Court of Federal Claims have all ruled against the ratification and wages theories, often in short opinions noting that the arguments have been rejected many times before. The Internal Revenue Service maintains a published list of frivolous positions, updated by notice, on which both the ratification theory and the wages theory appear, which puts filers on notice before penalties apply. Section 6702 imposes 5,000 dollars per frivolous submission, section 6702(b) reaches frivolous requests for hearings and installment agreements, and section 6673 authorizes the Tax Court to impose penalties for maintaining frivolous positions in litigation. Appellate courts add sanctions of their own for frivolous appeals. On the criminal side, Cheek holds that an honestly held misunderstanding can negate willfulness, but the misunderstanding must be genuine, and civil liability survives regardless.

There is a final reason these arguments will not be revived by a future court, and it is practical rather than doctrinal. A century of reliance now rests on the amendment: every revenue act since 1913, the entire administrative apparatus, and trillions of dollars of collections. Congress has reenacted the income tax repeatedly with full knowledge of the ratification challenges, which under standard interpretive principles confirms the earlier action even if a defect had existed. Courts give weight to such settled practice, and no judge has shown interest in unwinding it. The verdict stands as stated in the main section: false on the text, the certification, the decisions, and the uniform case law, with real penalties for pressing the theories in filings.

Claim two: paying income tax is voluntary

Stated in the strongest form its holders would recognize, this contention holds that the income levy is voluntary because the government’s own publications and officials describe the American system as one of voluntary compliance. If compliance is voluntary, the reasoning goes, then payment is optional, and no one can be compelled to file a return or remit money. Some versions add that the absence of a single statute saying in plain words that a named individual must pay proves the point.

The confusion here is more interesting than in the ratification claim, because the government’s language genuinely invites it. Internal Revenue Service publications, congressional testimony, and judicial opinions have all used the word voluntary to describe the American system. The error lies in what the word modifies. It modifies the method of assessment, not the obligation. The United States uses self assessment: the taxpayer computes the liability, completes the return, and remits the amount, rather than waiting for the government to calculate the bill and send it. That administrative choice is what voluntary describes. It has never meant that the underlying duty is optional, and the courts have said so repeatedly.

The controlling judicial statement is Flora v. United States, 362 U.S. 145, 176 (1960), in which the Supreme Court wrote: “Our system of taxation is based upon voluntary assessment and payment, not upon distraint.” The sentence describes who performs the initial calculation, the taxpayer rather than the revenue service, not whether the calculation may be skipped. Lower courts confronting the optionality reading have rejected it without exception, explaining that Flora’s language describes who does the arithmetic, not whether the arithmetic must be done. The distinction matters because it dissolves the apparent contradiction: a system can rely on citizens to assess themselves and still punish those who refuse, just as an honor system library can fine borrowers who keep the books. The Tenth Circuit made the enforcement side explicit in United States v. Tedder, 787 F.2d 540, 542 (10th Cir. 1986): “Congress gave the Secretary … the power to enforce the income tax laws through involuntary collection.”

The statutes impose the duty in plain terms. Section 1 of the Code imposes the tax on taxable income. Section 6001 requires persons liable for tax to keep records. Section 6011(a) requires persons made liable for tax to make a return when required by regulation. Section 6012 requires returns from individuals whose gross income exceeds the exemption amount, with thresholds that capture nearly all earners. Section 6151 provides that the person required to file shall pay the tax at the time and place fixed for filing. Withholding under section 3402 collects much of the liability before the return is ever filed, which is why the optionality reading collides with the lived experience of every wage earner: the money leaves the paycheck by operation of law, not by invitation. Sections 6651 and 6654 add penalties and interest for failure to file, failure to pay, and underpayment of estimated tax. The enforcement apparatus, liens under section 6321, levies under section 6331, and criminal provisions beginning at section 7201, exists precisely because the duty is not elective.

Holders of the voluntary reading sometimes point to the absence of a statute that names them personally and commands payment. That demand misunderstands how legislation works. Statutes impose duties on classes of persons defined by conduct and circumstance, residents with earnings above a threshold, employers paying wages, corporations earning profits, not on named individuals. Section 1 taxes the taxable income of every individual; the generality is the point, not a loophole. Courts addressing this variant have explained that the Code need not mention a taxpayer by name to bind that taxpayer, just as the speed limit need not mention a driver by name to fine that driver.

The historical record also undercuts the claim. From the Civil War income levy through the 1913 statute and every revenue act since, Congress has legislated compulsion: filing requirements, withholding, penalties, and criminal sanctions. The phrase voluntary compliance entered official usage to describe the remarkable fact that most Americans compute and pay without direct government assessment, a feature administrators prized because direct assessment of hundreds of millions of returns would be administratively impossible. To read that praise of the system’s efficiency as a grant of permission to opt out is to mistake a compliment for a loophole.

There is a narrow sense in which the word voluntary is accurate, and fairness requires stating it. Nobody is drafted into earning taxable income. A person who earns nothing owes nothing, and choices about work, investment, and residence shape the liability the Code imposes. In that thin sense, exposure to the levy follows from voluntary economic activity. But that is true of every tax and every fine in the legal system, and it does not make payment optional once the earnings exist. The verdict on the claim as its holders state it, that remittance is a matter of choice, is false, decided by sections 1, 6001, 6011, 6012, and 6151, the withholding provisions, the penalty structure, and the uniform judicial rejection of the optionality reading, with Flora supplying the authoritative explanation of what voluntary actually modifies.

The self assessment system predates the modern Code. The Civil War income tax of 1862 relied on taxpayers to report, assisted by assessors, and the 1913 statute did the same. The system’s scale changed in World War II: the Revenue Act of 1942 swept millions of new filers into the tax, and Congress needed collection to keep pace with the broader base. Congress answered with withholding on wages, which moved collection to the moment of payment, along with quarterly estimated payments for the self employed. Withholding did not change the legal duty; it changed the plumbing. The estimated tax provisions, now in section 6654, extended the same logic to income without withholding, with statutory safe harbors that excuse penalties for filers who pay enough during the year.

The phrase voluntary compliance entered the official vocabulary mid century as administrators contrasted the American approach with systems of direct government assessment used elsewhere. Congressional hearings and agency publications used it to praise the fact that citizens computed their own liability, a feature that made mass taxation administratively feasible. Critics later lifted the phrase from that context and read it as a legal conclusion. The courts’ answer, repeated across decades, is that context determines meaning: voluntary modifies compliance through self assessment, and nothing in the phrase repeals sections 1, 6012, or 6151.

A second family of variants argues that the Code’s duties apply only to certain categories of persons, variously defined as federal employees, residents of the District of Columbia, or holders of government privileges. The Code defines person in section 7701(a)(1) to include individuals, and United States person in section 7701(a)(30) to include citizens and residents, with citizenship itself governed by the Fourteenth Amendment and the nationality statutes. Courts have rejected the restricted definitions uniformly, noting that the provisions cited define terms for specific purposes rather than limiting the scope of the tax. Like the ratification theories, these arguments mistake specialized statutory language for a general limitation.

A third family of variants misreads the Code’s source rules. Sections 861 through 865 classify income as from sources within or without the United States, a classification that matters for foreign tax credits, nonresident aliens, and similar questions. Holders of the argument read the list of domestic source items in section 861(a) as an exhaustive catalogue of taxable income and conclude that domestic wages, not appearing there in the form they expect, are excluded. The reading ignores the companion provisions, the regulations, and the elementary point that source rules allocate income among categories rather than defining gross income, which section 61 already does. The Tax Court and the courts of appeals have rejected the argument, and the Internal Revenue Service lists it among frivolous positions subject to the section 6702 penalty.

The practical backbone of the self assessment system is third party information reporting. Employers file Forms W-2, payers file Forms 1099, and sections 6041 and 6045 require reporting of payments above thresholds. The Internal Revenue Service matches these reports against returns, which is why the voluntary label coexists with high compliance: most earnings are visible to the government independently of what the taxpayer reports. Noncompliance concentrates where reporting is thinnest, cash businesses and similar areas, which is why enforcement resources concentrate there as well.

The enforcement structure confirms the compulsory character. Section 6651 imposes additions to tax for failure to file and failure to pay, accruing monthly. Section 7203 makes willful failure to file or pay a misdemeanor, section 7201 makes willful evasion a felony, and Spies v. United States, 317 U.S. 492 (1943), clarified the affirmative acts that distinguish evasion from mere failure. Civil collection runs through the lien of section 6321, which attaches to all property upon assessment and demand, and the levy of section 6331, which after ten days’ notice and demand authorizes collection by distraint and seizure. A system with criminal penalties, liens, and levies is not describing an optional contribution, whatever vocabulary its administrators use for the assessment method.

The show me the law challenge deserves a direct answer because it sounds reasonable. Holders demand a single sentence commanding them personally to pay, and treat its absence as proof of voluntariness. The answer is that legislation does not work that way and never has. Section 1 imposes tax on the taxable income of every individual; section 6012 describes who must file by income level and filing status; section 6151 ties payment to filing. General language binding classes defined by conduct is how duties are legislated, from speed limits to jury service. Courts addressing the challenge have explained that the Code’s generality is a feature of drafting, not a defect in authority, and have sustained penalties against those who pressed the point.

Withholding tables add their own confusion because they approximate rather than compute. Employers apply tables and formulas that estimate annual liability from each pay period in isolation, which is why two workers with identical annual earnings can have different amounts withheld: pay frequency, bonuses, and midyear job changes all distort the estimate. The return corrects every approximation at once, which is why large refunds and large balances due usually signal withholding mismatch rather than anything about the schedule’s fairness. Adjusting the withholding certificate is the remedy the system provides; fearing the bracket is not.

The alternative minimum tax offers a useful contrast in rate structure. Its two tiers, 26 and 28 percent in the period covered, apply to a broader base with fewer preferences, and its exemption phases out over a pay range, creating a hidden marginal bump inside the phaseout. Taxpayers who straddle the regular and minimum systems can face effective marginal percentages that differ from either schedule’s headline figures. The complexity is real, and it is why the minimum tax draws perennial criticism, but its mechanics confirm the general principle: every schedule in the Code applies its percentages to slices of pay, and no slice reaches backward to retax dollars already assessed.

Claim three: a higher bracket can cut take-home pay

Stated in the strongest form its holders would recognize, this contention holds that earning more can leave a worker poorer, because crossing into a higher bracket subjects the whole paycheck to the higher percentage. On this view, a raise that pushes pay over a threshold triggers the new rate on every dollar, so the after tax total falls. Versions of the worry surface every filing season, in workplace conversations about overtime, bonuses, and promotions, and they influence real behavior: people decline extra hours from fear of the bracket.

The claim is false, and the reason is structural. The individual income levy applies marginal percentages, meaning each tier of the schedule touches only the dollars that fall inside it. Pay below a threshold keeps the lower percentage no matter how far above the threshold total pay climbs. Only the dollars above the line face the higher figure. A raise therefore always increases after tax income under the bracket schedule itself, because the pre raise dollars are untouched and the new dollars, though taxed at the higher percentage, still leave the earner with the after tax remainder.

A concrete illustration using the 2013 schedule for a single filer makes the mechanics visible. The first 8,925 dollars of taxable income faced 10 percent, the next band to 36,250 dollars faced 15 percent, and the band above that to 87,850 dollars faced 25 percent, with the schedule continuing through 28, 33, and 35 percent to a top rate of 39.6 percent above 400,000 dollars, the top rate restored for 2013 by the American Taxpayer Relief Act of 2012. Take an earner with 35,000 dollars of taxable income who receives a raise to 38,000 dollars, crossing the 36,250 dollar line into the 25 percent tier. Under the folk theory, the entire 38,000 dollars would now face 25 percent. Under the actual schedule, the first 8,925 dollars still face 10 percent, the dollars from 8,925 to 36,250 still face 15 percent, and only the 1,750 dollars above 36,250 face 25 percent. The tax on the 3,000 dollar raise is 625 dollars, so the raise adds 3,000 dollars of pre tax pay and 2,375 dollars after tax. The earner is richer, not poorer, and the same arithmetic holds at every threshold in the schedule, from the bottom tier to the 39.6 percent top rate.

What share of a raise lands in the new bracket?

Only the dollars above the threshold. If the 25 percent tier begins at 36,250 dollars of pay, a raise from 35,000 to 38,000 dollars puts 1,250 dollars in the old tier and 1,750 dollars in the new one. Everything below the line keeps its old percentage. The bracket taxes the margin, never the whole paycheck.

The distinction between the marginal percentage and the average percentage dissolves most of the remaining confusion. The marginal figure is the percentage applied to the next dollar earned. The average, or effective, figure is total liability divided by total income, and it is always lower than the top marginal percentage reached, because the lower tiers pull the average down. An earner in the 25 percent tier does not pay 25 percent of earnings; the earner pays 25 percent on the top slice and less on everything beneath it. Published tables from the Internal Revenue Service Statistics of Income division show this plainly: average percentages for each income group sit well below the top marginal figure for that group. Confusing the two is the single most common arithmetic error in popular tax discussion, and it feeds both the bracket fear and several of the distributional claims examined later in this article.

Fairness requires acknowledging the grain of truth that keeps the folk theory alive. While the bracket schedule itself never reduces take-home pay, the broader system contains phaseouts and cliffs that can impose very high effective percentages on additional earnings. Deductions, credits, and benefits often phase out over pay ranges, and each dollar of additional earnings can simultaneously face the bracket percentage and shrink a benefit, producing combined marginal burdens well above the statutory tier. The earned income credit phases out over a range, creating effective percentages in that range that surprise many filers. Certain benefit programs historically featured cliffs where a dollar of additional earnings cost many dollars of benefits. Social Security’s earnings test, before the full retirement age, once reduced benefits for earnings above a threshold, though later legislation softened it. These are real features of the system, and economists who study work incentives take them seriously. But they are phaseout mechanics, not bracket mechanics, and lumping them together with the tier schedule misidentifies the cause. The bracket claim as stated, that the tier itself can make a raise a pay cut, remains false under every schedule Congress has enacted.

There is also a behavioral footnote worth recording. Because the folk theory is widespread, it changes conduct: workers decline overtime, small business owners defer billings, and bonus recipients brace for disappointment. Employers and payroll departments spend real effort explaining that the fear is misplaced. The misunderstanding is therefore not merely academic; it distorts labor supply decisions at the margin. Correcting it does not require advanced mathematics, only the one sentence that the schedule taxes the margin rather than the whole. Verdict: false as a statement about the bracket schedule, decided by the statutory structure of section 1 and confirmed by the arithmetic of every published rate table, with the honest qualification that phaseouts elsewhere in the system can create high effective burdens that deserve their own analysis.

The rate schedule has a history that explains both its shape and its mythology. The 1913 act applied rates from one to seven percent, touching only a small fraction of households. The Second World War pushed the top rate to ninety-four percent and pulled millions of new filers into the system, creating the broad-based income tax that persisted after the war. The Kennedy-era cuts of 1964 brought the top rate down to seventy percent, where it sat until the 1981 act cut it to fifty and the 1986 act to twenty-eight. The 1993 legislation raised the top rate to thirty-nine and six-tenths percent, the 2001 and 2003 laws lowered it to thirty-five, and the January 2013 legislation restored thirty-nine and six-tenths percent at the top. Through all of these changes, the marginal structure never varied: each rate applied only within its bracket, and no reform ever made a raise reduce take-home pay under the regular schedule.

Bracket creep deserves its own paragraph because it is the real phenomenon that the myth distorts. Before 1985, the bracket thresholds were fixed in nominal dollars, so inflation alone pushed filers into higher brackets even when their real incomes had not risen. A worker whose wages merely kept pace with prices could face a higher marginal rate and a higher effective rate without any real raise, a hidden tax increase that Congress never voted. The 1981 act provided for indexing of brackets, exemptions and the standard deduction beginning in 1985, which ended the automatic drift. The myth that a raise can cost money may be a folk memory of the pre-indexing era, when inflation did quietly raise burdens; but even then the mechanism was rising prices, not the arithmetic of the brackets. Since indexing, the thresholds move with inflation, and the guarantee holds in nominal and real terms alike.

A second illustration, with round numbers, may help where the first did not. Imagine a simplified schedule: ten percent on the first ten thousand dollars of taxable income, twenty percent on the next ten thousand, thirty percent above twenty thousand. A filer with nineteen thousand dollars of taxable income pays one thousand dollars on the first slice and one thousand eight hundred on the second, for a total of two thousand eight hundred and an effective rate under fifteen percent. If income rises to twenty-one thousand, the extra two thousand dollars is taxed at twenty percent on the first thousand and thirty percent on the second, adding five hundred dollars of tax. Take-home income rises by one thousand five hundred dollars. At no point does the higher rate reach back into the lower slices, because the statute does not work that way; the brackets are containers, and each dollar is taxed according to the container it falls in.

A second concrete illustration, using a married couple filing jointly in 2013, shows the same pattern at a higher income. With 100,000 dollars of taxable income, the couple paid 1,785 dollars at the 10 percent tier, 8,197 dollars at the 15 percent tier, and 6,874 dollars at the 25 percent tier, for a total of 16,856 dollars. That is an effective rate of about 16.9 percent, not 25 percent of 100,000 dollars. Earning one more dollar would have added 25 cents of tax and 75 cents of take-home pay. The higher percentage never reaches backward into the lower slices, no matter how large the income grows.

The interaction with payroll taxes adds a wrinkle worth naming, because it is the one place where a raise can feel disappointing without contradicting the verdict. Payroll taxes apply from the first dollar of wages, so a worker whose income tax bracket does not change still sees the combined marginal rate exceed the income tax rate alone. And the wage cap means the payroll rate drops to zero above the maximum, which is why the combined marginal schedule has a hump shape for wage earners. None of this makes the income tax bracket reduce take-home pay; it means the total burden on a raise is the sum of several levies, each computed on its own base. The four-question sort applies again: which tax is being discussed determines which rate schedule answers the question.

Claim four: the wealthy pay nothing, and the wealthy pay everything

Two opposite contentions dominate popular discussion of distribution, and the strongest form of each deserves a hearing before either is graded. The first holds that affluent households pay no tax, or next to none, because deductions, preferential percentages on investment income, and sophisticated planning wipe out their liability. The second holds that affluent households pay essentially all the tax, because the income levy is steeply progressive and the bottom half of earners owe nothing. Both cannot be true as stated, and in fact neither survives contact with the published data intact. What survives is a subtler picture that depends entirely on the four questions from the opening section: which levy, which income measure, which unit, and which year.

Start with the income levy alone, because that is the levy both claims usually mean. The Internal Revenue Service Statistics of Income division publishes annual tabulations of individual returns by income group, and the pattern has been stable for decades. For tax year 2011, the latest data available on the article’s date, the top 1 percent of returns reported 18.7 percent of adjusted gross income and paid 35.1 percent of all individual income tax collected. The top 5 percent paid 56.5 percent of the income tax, and the top 10 percent paid 68.3 percent, while the bottom half of returns combined paid 2.9 percent. Thresholds move with the economy, but the shape persists across years: a small number of high income returns accounts for a large share of income tax receipts, and a large number of lower income returns accounts for a small share. Against this record, the claim that the wealthy pay no income tax is false. Against the same record, the claim that the wealthy pay all of it is an exaggeration that understates the contributions of the broad middle, which supplies roughly a third of income tax receipts.

Now widen the lens to all federal levies, which is the second question the four-question sort demands. The Congressional Budget Office publishes the standard distributional analysis, combining individual income, payroll, corporate, and excise levies and assigning each to households by income quintile. The edition current on the article’s date was the July 2012 release, and its framework is worth understanding because later editions revised some details. The office starts from market income, adds government transfers to get before tax income, subtracts federal levies to get after tax income, and adjusts for household size so that a single adult and a family of four are compared sensibly. Households are then ranked into quintiles, with the top quintile often subdivided to show the top 10, 5, and 1 percent separately. Corporate levies are allocated mostly to capital income, payroll levies to labor earnings, and excise levies to consumption. Each choice can be debated, and the office publishes its assumptions, but the framework is transparent and consistent across years, which makes trends readable. The consistent finding across editions is that average federal rates rise with income: the income levy dominates progressivity, rising most sharply with income, while the payroll levy works in the opposite direction at the top, because it applies only to earnings below the annual wage base, 113,700 dollars in 2013, and not at all to investment income, which makes its burden as a share of income fall for the highest earners. Excise levies, on fuel, tobacco, alcohol, and similar goods, are also regressive in the technical sense that lower income households spend larger shares of income on the taxed goods. None of this erases the progressivity of the whole, but it flattens it considerably relative to the income levy viewed alone.

The third measure is the effective percentage, total federal levies divided by income, which answers the question most people actually mean when they ask whether the system treats the affluent lightly or harshly. The distributional analyses show average federal percentages rising with income across the distribution, with the lowest quintile near zero, reflecting the combined effect of the earned income credit and the child credit, which can produce negative income tax liability, offset partly by payroll and excise burdens. These are averages across large groups, and individual households vary enormously around them: a high earner with large deductions or tax preferred investment income can face a far lower effective percentage than the group average, while a high earner with ordinary wage income faces a higher one. That variation is the factual seed from which the pay nothing claim grows. Published anecdotes about prominent investors paying lower effective percentages than their secretaries describe real arithmetic for those individuals, but they describe outliers, not the distribution. The group data show effective percentages rising with income, which is the definition of a progressive system, even as individual cases scatter widely.

Why does the unit of analysis change the answer?

Because households, tax units, and individuals sort people differently. Two earners filing jointly form one tax unit but may head one household or two. Quintile cutoffs move with the unit chosen, so the same dollars produce different shares at the top. Naming the unit before quoting a share prevents most misunderstandings.

The time period question adds a further complication that honest analysis must acknowledge. Annual snapshots classify people by one year’s income, which mixes the temporarily prosperous with the durably affluent: a business owner with one exceptional year lands in the top group alongside a career executive, though their lifetime positions differ. Lifecycle analysis, which follows people across decades, generally shows less inequality and less concentration of burden than annual snapshots, because many households move between quintiles over a working life. Neither measure is wrong; they answer different questions. But a claim about who pays what should name its period, because the annual and lifetime pictures differ enough to change the rhetoric built on them.

This is also the place to acknowledge the genuinely unresolved academic dispute over top income shares, the fraction of national income accruing to the highest earners, which matters because every distributional claim downstream depends on the income measure chosen. As the debate stood in September 2013, the live exchange ran between the top-income series of Thomas Piketty and Emmanuel Saez, built from individual tax returns and published in the Quarterly Journal of Economics in 2003 with annual updates thereafter, including Saez’s “Striking it Richer” update covering 2012 released in September 2013, which showed high and rising top shares of pre tax income, and critics who argued that changes in the tax base over time distorted the trend. The leading critical statement of that period came from Gerald Auten with Geoffrey Gee and Nicholas Turner, whose “New Perspectives on Income Mobility and Inequality” in the National Tax Journal in 2013 rebuilt the series with adjustments for base changes and income missing from returns, finding a flatter trend. The contested decisions include the unit of analysis, tax units versus individual adults; the income concept, pre tax market income versus income after transfers and taxes; and the allocation of income that never appears on individual returns, especially corporate retained earnings. Both camps publish methods and data, and the profession had not converged. A note on the brief’s anchor is required here: the exchange most associated with this dispute in later debate, between Emmanuel Saez and Gabriel Zucman on one side and Gerald Auten and David Splinter on the other, had not yet occurred on the article’s date. The Saez and Zucman paper first circulated as a working paper in October 2014 and appeared in the Quarterly Journal of Economics in May 2016, and the Auten and Splinter paper first circulated in November 2017. They are reported here as explicitly dated later data points, not as part of the 2013 debate, and they reproduce the same structure: one side finding high and rising top shares, the other finding lower and flatter ones after different technical choices, with no professional convergence. Any distributional claim that depends on a particular top share series should therefore be read with the series’ assumptions in view.

Grading the two original contentions against this record: the claim that the wealthy pay no tax is false, contradicted by every published tabulation, from the Statistics of Income division’s income tax shares to the Congressional Budget Office’s all federal analysis. The claim that the wealthy pay all the tax is partly true: affluent households do pay the dominant share of the individual income levy and a large majority of all federal levies, but the bottom four quintiles combined still supply a substantial share of the federal total, and the payroll levy in particular draws heavily from wage earners below the cap. Both claims, in their strongest forms, overstate. The data support progressivity with meaningful contributions across the distribution, a flatter picture than the income levy alone suggests, and an unresolved academic argument about exactly how concentrated top incomes are. That is less satisfying than either slogan, and it has the advantage of being what the record shows.

The most heated distributional argument of the period around 2011 and 2012 concerned the Tax Policy Center’s estimate that 46 percent of households would owe no federal income tax for 2011. The figure was accurate on its terms and widely misread. It reflected the recession’s toll on incomes, the standard deduction and personal exemptions sheltering low earnings, and refundable credits, especially the earned income credit and the child credit, wiping out liability for many working households. Its composition mattered: a large share were elderly living on Social Security, students, and low wage workers, groups with little income tax capacity in any system. And the same households paid payroll levies on their wages and excise levies on their purchases. The statistic described the income levy, not the federal system, which is the first of the four questions doing its work.

Which income measure is used changes the distribution picture as much as which levy is counted. Adjusted gross income is a tax concept, reduced by above the line deductions and excluding items like unrealized gains and much retirement saving. Cash income, used by the Tax Policy Center, adds back many of those items for a fuller pre tax picture. The Haig-Simons definition favored by economists, consumption plus the change in net worth, would include unrealized capital gains that never appear on a return. National income, used in the long run series built from tax data, allocates corporate retained earnings and other non filer income across households by assumption. Top income shares look highest under the broadest measures and flattens under after tax, after transfer measures. None is the uniquely correct choice; each answers a different question.

The time period question adds the lifetime perspective. Annual snapshots classify a medical resident and a retired executive by one year’s income, though their lifetime positions differ enormously. A Treasury study tracking filers from 1996 to 2005 found substantial movement: more than half of those in the bottom quintile in 1996 had moved to a higher quintile by 2005, and movement down from the top occurred as well. Lifetime incidence studies generally find the system less progressive over a lifetime than in any single year, because transfers received early and taxes paid later partly offset. The annual and lifetime views are complements, not competitors, but rhetoric built on one while ignoring the other misleads.

A final application of the which tax question: state and local levies look very different from federal ones. Studies of state and local systems consistently find them regressive overall, because they rely heavily on sales and property levies that claim larger shares of low incomes. A complete picture of an American household’s total tax burden must combine the progressive federal system with the regressive tilt of most state and local systems. Claims about the total burden that cite only federal figures understate what lower income households pay.

Income concentration and wealth concentration are related but distinct, and confusing them feeds both slogans. Wealth is far more concentrated than annual income: the Federal Reserve’s Survey of Consumer Finances for 2010 showed the top 10 percent of households holding roughly three quarters of net worth, a much steeper gradient than any income series. But wealth is not taxed annually by the federal government, except through the estate levy on a small fraction of decedents, so wealth concentration does not translate directly into claims about who pays the income tax. Keeping the stock of wealth and the flow of income analytically separate prevents a whole class of errors.

The estate and gift taxes, though small as a share of receipts, generate myths of their own that the same method dispatches quickly. The claim that the estate tax destroys family farms and small businesses has circulated for decades, yet the published data show the tax reaching only a tiny fraction of estates, with special valuation and deferral provisions available for qualifying farm and business assets. The exemption stood at 5.25 million dollars per person in 2013, which placed the vast majority of estates outside the tax entirely. Conversely, the claim that the estate tax is painless because it touches so few ignores the planning costs and the genuine liquidity problems of asset-rich, cash-poor estates, which the deferral provisions only partly solve. Both claims take a true feature, the narrow reach or the real compliance burden, and inflate it into a general description. The verdicts are the familiar pair: false as stated, informative once qualified, and best read with the statute’s actual thresholds in hand.

State corporate levies add a layer the federal debate often omits. Most states impose their own corporate income taxes, with percentages generally between 4 and 10 percent, and the states have developed elaborate doctrines, combined reporting, throwback rules, and nexus standards, to apportion multistate income. A company’s total government burden is the federal levy plus these state levies, and planning that reduces the federal figure does not always reduce the state one. Competitiveness comparisons that cite only the federal headline understate the combined burden, while effective rate studies that include state levies show a fuller picture.

The inversion wave of the period illustrated the stakes of the international rules. Beginning in the 1990s and accelerating through the early 2010s, a number of American companies restructured so that a foreign parent sat atop the group, aiming to escape the domestic levy on foreign earnings and to strip domestic income through interest payments to the new parent. Congress responded in 2004 with section 7874, which treats inverted companies as domestic when the former American shareholders retain sufficient ownership and the group lacks substantial foreign business. The provision slowed but did not end the practice, and each subsequent transaction renewed the policy argument. The episode belongs in any honest account of corporate taxation because it shows the system responding to planning in real time, through legislation rather than through the folk theory that nothing is collected.

Claim five: corporations pay no tax

Stated in the strongest form its holders would recognize, this contention holds that the corporate income levy is a fiction: large companies use deductions, credits, offshore arrangements, and lobbying to reduce their liability to zero, so the statutory percentage is theater and the government collects nothing from the business sector. A mirror image version holds that the corporate levy falls entirely on shareholders and therefore does not touch workers or consumers at all. Both versions overclaim, and the honest answer requires separating three questions the four-question sort keeps distinct: what the government collects, what individual companies pay, and who ultimately bears the burden.

Start with collections. The corporate income levy is a substantial but secondary source of federal receipts. In fiscal year 2012, the levy produced about 242 billion dollars, roughly 10 percent of total federal receipts, a share that has moved within a band of about 7 to 13 percent across recent decades. The individual income levy and the payroll levies each supply roughly four times that amount. The corporate share was larger in the mid twentieth century, exceeding 25 percent of receipts in some years of the 1950s and 1960s, and it has trended down as a fraction of both receipts and national output. It has never been zero in the aggregate, and no fiscal year on record shows the business sector escaping the levy entirely.

Turn to the statutory and effective percentages, where the claim finds its most sympathetic facts. The federal statutory percentage stood at 35 percent through 2013, and combined with state levies the headline figure approached 39 percent, among the highest in the industrialized world at the time. But the statutory figure is the beginning of the calculation, not the end. Deductions for the costs of earning income, credits enacted for research, investment, and other favored activities, loss carryforwards that let bad years offset good ones, and deferral of tax on foreign earnings until repatriation all reduce the effective percentage below the headline. Studies of effective percentages have found that many large corporations pay effective percentages well below the statutory figure and that a substantial fraction of large corporations report no federal liability at all in particular years. A 2008 study by the Government Accountability Office, GAO-08-957, found that about 55 percent of large corporations under American control reported no federal tax liability in at least one year between 1998 and 2005. Those zeros are real, and they reflect the interaction of losses, credits, and timing provisions rather than lawbreaking. But a zero in one year for one company is not a zero for the sector across time, and profitable companies without large credits or losses pay substantial amounts.

The international dimension deserves a plain statement because it feeds much of the rhetoric. American companies operating abroad could defer the domestic levy on foreign earnings until those earnings were brought home, which encouraged the accumulation of profits in foreign subsidiaries and elaborate planning around the location of income. Whether particular arrangements constituted avoidance or legitimate planning under the Code’s own rules was often contested, and Congress revisited the boundary repeatedly. What the record supports is that deferral was a deliberate feature of the pre 2018 system, not a loophole discovered by accident, and that its revenue cost was one reason later reform debates focused on international rules. Describing lawful use of deferral as paying no tax confuses a single company’s planning outcome with the sector’s aggregate contribution.

The deepest difficulty, and the one that earns the unresolved label, is incidence: who actually bears the corporate levy once markets adjust. A corporation writes the check, but the burden may land on shareholders through lower returns, on workers through lower wages, or on consumers through higher prices, depending on how capital and labor respond. Economists genuinely disagree, and the disagreement is methodological rather than ideological. The classic closed economy analysis, associated with Arnold Harberger’s 1962 work, found that the levy falls largely on owners of capital, because capital cannot easily flee the domestic economy. Open economy models, developed by Harberger himself in later work and by others studying capital mobility across borders, find that in a world where capital moves freely, the levy can depress domestic investment and wages, shifting much of the burden to labor. Empirical estimates scatter across a wide range.

The official scorekeepers have taken positions while acknowledging the uncertainty. The Congressional Budget Office, in the distributional analyses current in the period, allocated 75 percent of the corporate levy to owners of capital and 25 percent to labor income, a change from its earlier practice of attributing the entire levy to capital owners. The Treasury Department’s Office of Tax Analysis, in Technical Paper 5 of May 2012, used an 82 percent capital and 18 percent labor split. The Joint Committee on Taxation used a 75 percent capital and 25 percent labor split from 2013 onward, with 100 percent to capital in the short run. All three describe these as assumptions for presentation, not findings of settled science, and all have revised their assumptions before. The academic literature spans the range: a March 2013 review by Jennifer Gravelle in the National Tax Journal surveyed open economy general equilibrium estimates and found that, adjusting to central empirical elasticity estimates, capital bears the majority, about 60 percent, of the burden, while stressing that the results vary widely with elasticity and modeling assumptions, with some models assigning the majority to labor. The honest statement is that the incidence of the corporate income levy is not known with confidence, that credible economists defend positions across the spectrum, and that any distributional claim that depends on a particular incidence assumption inherits that uncertainty. Note the author: the 2013 review is by Jennifer Gravelle of the Congressional Budget Office, not to be confused with the Congressional Research Service analyst of a similar name who wrote the CRS corporate tax reports.

Grading the original contention: the blanket claim that corporations pay no tax is false at the aggregate level, contradicted by hundreds of billions in annual collections, and misleading at the company level, where zeros in particular years reflect the Code’s own loss, credit, and timing provisions rather than absence of the levy. The narrower observation that many large companies pay little or nothing in particular years is true and documented. The mirror claim about who bears the burden is genuinely unresolved, with the Congressional Budget Office, the Treasury Department, and the Joint Committee on Taxation publishing different allocation assumptions and the academic literature spanning the full range from capital to labor. Here the four-question sort does its most useful work: which companies, which years, and whose burden must all be named before the claim can even be evaluated.

The corporate levy’s history is a long decline in the headline percentage from wartime peaks. An excise of 1 percent on corporate income above 5,000 dollars arrived in 1909, the 1913 act set 1 percent, and wartime excess profits levies pushed the combined burden far higher in the 1940s and 1950s, when the statutory figure stood at 52 percent. Reductions followed: 48 percent in the 1964 act, 46 percent in 1978, 34 percent in the 1986 reform, and 35 percent from the 1993 reconciliation act, where it stood through 2013. Combined with state levies averaging around 4 to 5 percent, the American headline figure approached 39 percent, compared with an Organization for Economic Cooperation and Development average near 25 percent in 2012. The gap between the American statutory figure and the international average was the starting point of every competitiveness argument of the period.

The effective percentage sits well below the headline because the Code deliberately narrows the base. Accelerated depreciation under the modified accelerated cost recovery system lets companies recover equipment costs faster than economic wear, reducing early year liability. The domestic production deduction in section 199, fully phased to 9 percent by 2010, cut the effective percentage on qualifying income by more than 3 points. The research credit in section 41 subsidizes experimentation. Interest on debt is deductible under section 163 while returns to equity are not, which favors debt finance and shrinks the base. Each provision was enacted for a stated policy reason, and each moves the effective percentage further from the statutory one. Studies of effective percentages therefore show wide variation by industry: capital intensive manufacturers using depreciation look very different from service firms.

The international rules supplied the most controversial base narrowing. Under the system in effect through 2013, American companies owed the domestic levy on worldwide income but could defer it on active foreign earnings until repatriation, while claiming credits for foreign taxes paid under section 901. Deferral was not an oversight; Congress designed it, fenced it with the anti deferral regime of subpart F in sections 951 through 964, and debated its scope for decades. The practical result was large accumulations of earnings in foreign subsidiaries and sophisticated planning around where income was booked. The 2004 American Jobs Creation Act offered a one time repatriation holiday at an effective 5.25 percent under section 965, which brought home hundreds of billions and demonstrated both the scale of the stockpile and the responsiveness of repatriation to the price. Whether deferral constituted a sensible competitiveness policy or an unjustified preference was contested throughout the period.

Underneath the planning debate sits the older question of double taxation. The classical system taxes corporate profits at the company level and dividends again at the shareholder level, which critics argue distorts choices toward debt and retained earnings. The 2003 act’s reduction of the dividend and capital gain percentages to 15 percent functioned as partial integration, narrowing the double burden without adopting a full imputation system. A Treasury study in 1992 had mapped integration options, from dividend exclusion to shareholder credits, and found each with tradeoffs in revenue and complexity. The double taxation complaint is therefore partly true as arithmetic and genuinely contested as policy, with reasonable analysts disagreeing about how much the distortion matters.

The gap between book income and taxable income explains much of the remaining confusion, and it deserves a plain description. Corporations report profits to shareholders under financial accounting rules and compute tax under the Code, and the two measures differ by design: depreciation schedules, stock option deductions, foreign earnings treatment and dozens of smaller provisions create a book-tax gap that researchers have tracked for decades. A firm can report billions in book profit while owing far less in tax without breaking any rule, because the two numbers answer different questions for different audiences. Critics cite the gap as evidence of avoidance; defenders cite it as evidence that Congress deliberately subsidizes investment through the tax system. Both readings describe the same arithmetic. The gap is a feature of a Code that uses tax provisions to pursue non-revenue goals, and judging it requires deciding whether those goals are worth the complexity, which is a policy argument rather than a myth to be graded true or false.

The zero liability years that feed the folk claim deserve a precise accounting. Net operating losses under section 172 let companies carry losses forward to offset later profits, so a company recovering from a bad year can owe nothing while profitable on the year’s operations. General business credits, including the research credit and energy credits, directly reduce liability. Timing differences between book and tax accounting, especially depreciation, shift liability across years. None of these is evasion; all are the Code operating as written. The Government Accountability Office’s studies documenting the frequency of zero liability years among large corporations describe the interaction of these provisions, not a failure to enforce the law.

Small corporations add a coda. S corporations and partnerships pass income through to owners taxed under the individual provisions, so much of American business income never faces the corporate levy at all. The growth of pass through taxation since the 1986 reform, which for a time left the top individual percentage below the corporate one, shifted activity out of C corporation form. Claims about business taxation that ignore pass throughs miss most small firms entirely, which is one more reason the which tax and which unit questions from the opening section cannot be skipped.

A related hypothesis holds that deficits are themselves the strategy: cut taxes, let the resulting shortfalls force spending restraint, and accept the revenue loss as the price of smaller government. The idea circulated in policy debate under the shorthand starve the beast. Its empirical record is mixed at best. The 1980s combined large tax cuts with large spending increases, particularly defense, and the 2000s combined cuts with spending growth in defense, entitlements, and domestic programs. Deficits rose in both decades without producing the predicted retrenchment. Whether the hypothesis fails as politics or as economics, the revenue record shows the cuts did not pay for themselves first, which is the only claim this article grades.

The long run budget outlook published in 2013 put the scoring debates in structural context. The Congressional Budget Office’s extended baseline showed revenues recovering toward historical averages as the economy healed, while spending on health programs and retirement benefits drove deficits wider over subsequent decades under unchanged law. In that framework, rate cuts that lose revenue relative to baseline deepen an already challenging path, regardless of their incentive effects. The outlook did not take a position on the right size of government; it showed the arithmetic that any position must satisfy.

Claim six: the 1981 or 2017 rate cuts paid for themselves

Stated in the strongest form its holders would recognize, this contention holds that cutting tax percentages increases economic activity so much that the government collects as much or more revenue afterward as before, making the cuts self financing. The claim attaches most famously to the Economic Recovery Tax Act of 1981, and in later debate to subsequent rate reductions. Its holders point to rising nominal receipts after the cuts and argue that the growth proves the point. Its skeptics point to the official revenue estimates published before passage and to the deficits that followed. The record that decides between these readings is the published scoring, and this section reports it rather than characterizing it.

The 1981 act cut individual percentages substantially, reducing the top marginal figure from 70 percent to 50 percent and cutting lower tiers as well, alongside business provisions including accelerated depreciation. Before passage, the staff estimate published in Senate Report 97-144 put the revenue effect at a loss of 688.1 billion dollars over fiscal years 1981 through 1986. That estimate was static: the report stated plainly that it did not include assumptions about taxpayers changing their behavior in response to the cuts and did not take feedback effects into account. It represented the official score for the Finance Committee bill, produced by the estimators Congress created for exactly this purpose. The Congressional Budget Office’s budget projections told a consistent story, showing deficits widening as the revenue loss took effect.

What happened afterward is a matter of published fiscal history. Federal receipts, which had stood at 19.6 percent of gross domestic product in fiscal 1981, fell to 17.3 percent by fiscal 1984 before recovering partially later in the decade. In nominal dollars receipts eventually grew, because the economy grew and inflation raised nominal incomes, but as a share of the economy they ran below their pre cut level for years. The deficit, which had been 2.6 percent of gross domestic product in fiscal 1981, topped 5 percent in the mid 1980s, and the national debt roughly tripled in nominal terms across the decade. None of these figures is in dispute; they appear in the historical tables the Office of Management and Budget and the Congressional Budget Office publish each year.

The analytical question is how much of the revenue path to attribute to the cuts versus everything else happening in the economy: the deep recession of 1981 to 1982, the Federal Reserve’s disinflation, the subsequent recovery, later tax legislation including the 1982 and 1984 deficit reduction acts that clawed back a portion of the 1981 cuts, and the 1986 reform discussed below. Professional estimates of the feedback effect, the additional revenue generated because lower percentages encourage more economic activity, have consistently found feedback offsetting only a fraction of the static loss. Analyses published by the Congressional Budget Office, the Joint Committee on Taxation, and the Treasury Department across the 1980s and afterward modeled behavioral responses, including changes in labor supply, saving, and investment, and found real but partial offsets. The curve illustrating that a zero percent rate and a 100 percent rate both yield zero revenue, popularized in policy debate, implies that some rate maximizes collections, but it does not imply that the pre 1981 rates sat above that maximum, and the published estimates concluded they did not.

The Tax Reform Act of 1986 provides a useful contrast because it was designed under an explicit revenue neutrality constraint: the Joint Committee on Taxation scored its rate reductions and base broadening provisions as offsetting each other within the budget window. The record of behavioral response evidence from the Tax Reform Act of 1986 shows how economists studied taxpayer reactions to that reform, including the well documented shifting of income across the reform’s effective date, which is a reminder that measured responses to rate changes include timing shifts that do not represent lasting changes in economic activity. The 1986 experience demonstrates that Congress knew how to enact rate cuts that the official scores showed as paid for, by pairing them with offsetting provisions, which makes the absence of such offsets in 1981 analytically significant rather than incidental.

How should a reader read an official revenue estimate?

Ask three things: whose number it is, which method they used, and what baseline they measured against. A conventional estimate holds the economy fixed and counts the mechanical effect plus microeconomic responses. A dynamic estimate adds macroeconomic feedback. An estimate that names all three can be evaluated; a slogan that names none cannot.

For readers who want the mechanics of how these estimates are produced, the explanation of how tax bills move through Congress describes the scoring process: the Joint Committee on Taxation estimates revenue effects against a current law baseline, the Congressional Budget Office projects the budgetary path, and both sets of numbers discipline the legislative debate. The estimates are projections, not prophecies, and they are revised as data arrive, but they are the profession’s best contemporaneous judgments, published openly and subject to after the fact evaluation.

The fuller account of the 1981 legislation, including its provisions and its legislative history, appears in the guide to the Economic Recovery Tax Act of 1981, and the later round of rate reductions is treated in the guide to the Tax Cuts and Jobs Act of 2017, which carries the modern revenue estimates and official scores for that legislation. The pattern relevant to the claim is that each round generated the same contention, that each round was scored by the same nonpartisan estimators, and that the published scores in each case projected net revenue losses after accounting for modeled economic feedback. Reporting those scores is not a characterization of the claim’s holders; it is the record.

The 2017 law’s score belongs here as an explicitly dated later data point, not as knowledge available in 2013. The Joint Committee on Taxation’s final conference committee estimate, JCX-67-17 of December 18, 2017, projected that the law would reduce revenues by 1.455 trillion dollars over fiscal years 2018 through 2027 on a conventional basis. The Committee’s dynamic analysis, adding macroeconomic feedback from higher investment and output, estimated that growth effects would offset roughly a third of the conventional loss. No published score showed the law financing itself within the budget window. The core point both scores support is the same one the 1981 score supports: neither round of cuts was officially scored as paying for itself.

What does paying for itself mean in official scoring?

It means added economic activity restores the receipts a rate cut removes, within the budget window, leaving the deficit unchanged. Scorekeepers model that feedback and publish the net figure. When the published net still shows a loss, the cut did not pay for itself on the scorekeepers’ own terms.

Verdict on the 1981 version: the claim that the cuts paid for themselves is contradicted by the official scores published before passage and by the fiscal history published afterward, which showed receipts falling as a share of the economy and deficits widening. The honest qualification is that feedback effects are real and were included in the estimates, so the claim fails on the scorekeepers’ own terms rather than on terms its holders were never offered. On the recurring form of the claim as applied to later cuts, the same standard applies: consult the published estimates, which for the 2017 round are carried in the series guide linked above.

The 1981 act did not survive intact, which complicates any simple story about its effects. The Tax Equity and Fiscal Responsibility Act of 1982, the largest peacetime revenue increase to that point, reversed a substantial fraction of the 1981 cut over subsequent years, and the Deficit Reduction Act of 1984 took back more. Later analyses of the decade’s fiscal path therefore measure the combined effect of the cut and its partial reversals, not the 1981 act alone. The 1986 reform then restructured the system under a revenue neutrality constraint, trading lower percentages for a broader base. Attributing the decade’s revenue outcome to the 1981 act in isolation, in either direction, ignores the legislation Congress passed to modify it.

The 1990s supply the most cited counterpoint. The 1990 and 1993 reconciliation acts raised top percentages, and the Joint Committee on Taxation scored both as revenue gaining. Receipts then climbed from under 18 percent of gross domestic product early in the decade to 20.6 percent by fiscal 2000, and the budget ran surpluses from 1998 through 2001. Advocates of the self financing view attribute the surge to the earlier cuts’ growth effects; skeptics attribute it to the rate increases, the technology boom, capital gains realizations, and bracket creep as inflation pushed earners into higher tiers. The professional literature treats it as overdetermined: many causes operated at once, which is precisely why controlled scoring, rather than after the fact storytelling, is the better evidence.

The 2001 and 2003 cuts replayed the analytical exercise with newer tools. The Joint Committee on Taxation scored the 2001 act at about 1.35 trillion dollars of revenue loss over ten years, and the 2003 act’s dividend and capital gain reductions added further scored losses. Receipts fell from 20.6 percent of output in fiscal 2000 to about 16 percent in fiscal 2004, and deficits returned. Defenders noted the 2001 recession and the attacks of that year as confounding events, which is fair, and also fair is that the estimators’ models included the expected growth feedback and still showed large net losses. The Congressional Budget Office’s retrospective analyses found the revenue path broadly consistent with the scored losses plus the cycle, not with self financing.

Understanding the baseline matters because much confusion hides there. Official scores measure against current law, meaning the revenue path if Congress did nothing, not against last year’s collections. A cut can therefore coincide with rising nominal receipts, because the economy grows, while still losing revenue relative to the baseline. Holders of the self financing view often cite the nominal rise; the estimators cite the baseline gap. Both statements can be factually accurate and still disagree, which is the four-question sort’s time period and measurement questions resolving another apparent contradiction.

The scoring technology itself became a policy battlefield. Traditional estimates held the size of the economy fixed and measured only the mechanical revenue effect plus microeconomic behavioral responses, such as timing shifts. Dynamic scoring adds macroeconomic feedback: lower percentages may expand the economy, enlarging the base. The Joint Committee on Taxation published macroeconomic analyses of major proposals in 2003 and 2005, the Treasury’s Office of Tax Analysis built dynamic models used in 2006, and the Congressional Budget Office produced dynamic analyses of tax policy across the decade. The consistent finding was that feedback offsets a fraction of the static loss for rate cuts, larger for cuts aimed at investment, smaller for cuts aimed at consumption, but not the whole. Proponents of fuller dynamic scoring argued the models understated incentive effects; the estimators answered that the models already included them and the question was magnitude, not existence.

The underlying theory deserves a fair statement. Lower marginal percentages do improve incentives to work, save, and invest at the margin, and the empirical literature finds real responses, especially in the timing of income and in reported taxable income among high earners. The classic estimate is Martin Feldstein’s 1995 study of the 1986 reform in the Journal of Political Economy, volume 103, pages 551 to 572, which used a Treasury panel of more than 4,000 taxpayers and estimated the elasticity of taxable income with respect to the net of tax rate at one or higher, the high end of the literature, though later work found the estimate overstated because of mean reversion. The critical survey by Emmanuel Saez, Joel Slemrod, and Seth Giertz in the Journal of Economic Literature, volume 50, 2012, pages 3 to 50, found a wide range of estimates across studies, generally lower for the 1990s reforms than for the 1980s, and concluded that the elasticity is higher at top incomes where avoidance opportunities are greater. Behavioral response to rates is real; its magnitude is contested, which is exactly the kind of qualified finding a myths article should report. The 1986 reform’s well documented shifting of income into 1986 ahead of 1987’s lower percentages showed taxpayers responding sharply to anticipated changes. But incentive effects and revenue feedback are different magnitudes: an economy can grow faster while still yielding less revenue than the higher rate path would have, because the broader base does not fully replace the lower percentage. Growth and self financing are related but distinct claims, and the evidence supports the first more readily than the second.

The standard this article applies is therefore the one the institutions themselves use: the published estimate, produced before passage, incorporating the modeled feedback, measured against the baseline of unchanged law. On that standard, the 1981 cuts were scored as large net revenue losses, the 2001 and 2003 cuts likewise, and the fiscal histories tracked the scores more closely than the self financing contention. The 2017 round’s estimates are carried in the series guide linked above, where the same standard can be applied.

Academic work using different methods reached compatible conclusions about the direction of effects. Christina Romer and David Romer’s 2010 study used the narrative record of legislated tax changes to identify exogenous shifts and found that tax increases are followed by reduced output, with an implied multiplier suggesting real economic costs to raising revenue. Their work is often cited by advocates of lower rates, and fairly so, but its implications for the self financing question run the other way: finding that taxes affect output is not finding that cuts pay for themselves. The Romers’ estimates imply that the revenue feedback from a cut offsets only part of the static loss, consistent with the official scores. Distinguishing the growth question from the revenue question is the analytical move the whole debate requires, and the literature supports it.

Claim seven: payroll taxes sit in a personal account

Stated in the strongest form its holders would recognize, this contention holds that the payroll levies deducted from each paycheck are deposited into a personal account bearing the worker’s name, where the money accumulates until retirement, much like a savings account or a private pension. On this view, benefits in old age are simply withdrawals of one’s own money, the government acts as custodian of individual balances, and any shortfall in the system reflects mismanagement of those balances.

The contention is false, and the statute explains why in its first provisions on the subject. The Federal Insurance Contributions Act, codified at sections 3101 and 3111 of the Internal Revenue Code, imposes the levies: the employee portion on wages under section 3101 and the matching employer portion under section 3111. The money collected does not enter individual accounts. Section 401(a) of title 42 of the United States Code creates the Federal Old-Age and Survivors Insurance Trust Fund on the books of the Treasury and appropriates to it amounts equivalent to the payroll taxes collected. Current collections pay current beneficiaries, with surpluses in some years held as Treasury securities within the trust funds. The system is financed on a pay as you go basis, supplemented by the trust fund balances, and no provision of law creates a segregated account for any individual worker. The trust funds are government accounts established for internal accounting, which lets the government track revenues dedicated for specific purposes; they are not vaults of personal balances.

The Supreme Court settled the legal character of the arrangement more than half a century ago. In Flemming v. Nestor, 363 U.S. 603 (1960), the Court held that workers have no contractual or property right to benefits based on their contributions. Congress may alter benefit formulas, eligibility rules, and payment levels through legislation, and contributors cannot sue for the return of their payments as if the money were held in trust for them individually. The decision used direct language: the interest of a covered employee in future benefits is not a property interest protected against legislative change. Whatever political or moral claims contributors may feel they hold, the legal claim to a personal balance does not exist.

What the system does record for each worker is an earnings history, and the distinction between an earnings record and an account balance is the key to dissolving the confusion. The Social Security Administration maintains lifetime earnings records, and the benefit formula applies to those records: it averages indexed earnings over the working life, applies a progressive formula that replaces a larger share of earnings for lower earners, and produces a monthly benefit amount. The earnings record determines the size of the future benefit, but it does not represent money set aside. Two workers with identical earnings histories receive identical computed benefits even if one contributed during years of surplus and the other during years when the system drew on reserves, because the benefit depends on the record, not on any balance.

Why does no statement arrive showing a personal balance?

Because the statute creates no individual account to report. Contributions are recorded as earnings history, which sets future benefit levels, while the money itself flows into government trust funds and out to current beneficiaries. A statement would imply ownership of a balance the law never established.

The annual statements the administration mails to workers reinforce the confusion in a subtle way. They show estimated future benefits based on the earnings record, and readers naturally interpret an estimate tied to their own work history as a report on their own money. The estimate is genuine, but it describes a legislated benefit computed from the record, not a balance in an account. Congress can change, and has changed, the formulas, the retirement ages, and the taxation of benefits, which it could not do if the statements reported private property.

Several concrete features of the system are inexplicable on the personal account theory and natural on the actual design. The levy applies only to earnings below the annual wage base, 113,700 dollars in 2013, which would be an odd design for a savings account but makes sense as a policy choice about the scope of a social insurance program. The Medicare portion of the payroll levy, uncapped since 1993 legislation removed its wage base, finances health benefits unrelated to any individual’s contributions. Spouses and survivors can receive benefits based on a worker’s record without having contributed themselves. Disability benefits flow to workers who become disabled long before any plausible account could fund them. Each of these features reflects a program of social insurance with redistributive elements, not a system of individual savings.

The fuller history of how the program’s financing evolved, including the 1935 enactment, the 1939 amendments that added dependents and survivors, the 1983 amendments that addressed the financing shortfall then projected, and the introduction of benefit taxation, is carried in the history of the Social Security amendments, which belongs to the series’ retirement cluster and treats the trust fund mechanics at length. For present purposes the verdict is straightforward: the claim that payroll levies accumulate in a personal account is false, decided by the Federal Insurance Contributions Act’s trust fund structure, the absence of any statutory individual account, and Flemming v. Nestor’s holding that no property right attaches to contributions. The earnings record is real, the benefit formula is real, and the money in any individual worker’s name is not.

The program’s financing was pay as you go nearly from the start, though that was not the original vision. The 1935 act imagined building a large reserve from contributions, with benefits beginning later. The 1939 amendments transformed the design: they added benefits for dependents and survivors, moved the first monthly payments up to 1940, and converted the system into family insurance financed on a current cost basis. The reserve concept survived only in the trust funds’ accounting. From 1940 onward, the contributions of current workers paid the benefits of current retirees, and every subsequent financing debate has taken that structure as given.

The trust funds are government accounts, not vaults of cash. Surplus collections are credited to the funds as special issue Treasury securities, which earn interest and are redeemed when benefit payments exceed collections. Between 1984 and 2009 the funds ran surpluses nearly every year, building reserves that exceeded 2.5 trillion dollars. Since then the system has drawn on interest and then principal, exactly as the 1983 financing repair anticipated. Critics sometimes describe the securities as worthless IOUs; legally they are obligations of the Treasury like any other, backed by the same full faith and credit, and redeeming them requires the government to raise the cash through taxes or borrowing, which is a fiscal fact, not a sign of fraud.

The 1983 amendments, following the National Commission on Social Security Reform chaired by Alan Greenspan, were the largest financing repair in the program’s history. They accelerated scheduled rate increases, brought federal civilian employees into the system, began taxing a portion of benefits for higher income retirees, raised the full retirement age gradually from 65 to 67, and adjusted the benefit formula’s indexing. The package was designed to produce decades of surpluses followed by gradual reserve depletion, and the trustees’ projections at the time mapped that path. Whatever one’s view of the policy mix, the episode demonstrates the point Flemming made as law: Congress adjusts the program’s terms legislatively, because no contributor holds a contractual claim.

The benefit formula translates the earnings record into a monthly amount through deliberately progressive arithmetic. The administration indexes lifetime earnings, averages the highest 35 years to get average indexed monthly earnings, and applies the primary insurance amount formula: 90 percent of the first bend point, 32 percent of the next segment, and 15 percent above the second bend point, with the bend points at 791 and 4,768 dollars in 2013. The 90 percent bracket replaces most of a low earner’s wages; the 15 percent bracket replaces little of a high earner’s. Spouses can claim up to half the worker’s amount, survivors receive adjusted amounts, and disability insurance, added in 1956, pays workers who become severely disabled. Each feature is inexplicable as a savings account and natural as social insurance.

The retirement earnings test illustrates the program’s insurance character. Workers claiming benefits before the full retirement age who continue to earn above an exempt amount see benefits reduced, historically one dollar for every two or three above the line, with the withheld amounts later restoring higher benefits through recomputation. Legislation in 2000 eliminated the test above the full retirement age. A personal account would have no reason to penalize continued work; an insurance program balancing adequacy and cost does.

The Medicare hospital insurance portion, financed by the uncapped payroll levy since 1993, reinforces the analysis. Its trust fund faces its own financing outlook, its benefits bear no relation to any individual’s contributions, and general revenues openly subsidize the supplementary medical insurance trust fund. The payroll levy’s two parts together illustrate the design: earnings related contributions funding a mix of earned benefits and social insurance transfers, with no individual balances anywhere in the structure.

Disability insurance, the often forgotten half of the acronym, reinforces the point. The Disability Insurance trust fund, financed from the same payroll tax at a smaller allocation rate, pays benefits to workers who become disabled before retirement age, and its finances have run tighter than the retirement fund’s. A worker who never becomes disabled receives nothing from that portion of the tax, which is exactly what insurance means and exactly what a personal account would not do. The system’s defenders cite this risk pooling as its moral core; its critics cite the same pooling as evidence that the link between contributions and benefits is weaker than advertised. Both sides are describing the same institution. The myth that each worker has an account mistakes a social insurance program for a savings plan, and no feature of the law supports the mistake.

Beginning in 2013, an additional nine-tenths of one percent hospital insurance tax applied to wages above two hundred thousand dollars for single filers and two hundred fifty thousand for joint filers, and a three and eight-tenths percent tax applied to certain investment income above the same thresholds. These provisions, enacted as part of the 2010 health legislation, made the payroll side of the system modestly more progressive at the top, because the additional tax has no wage cap. A claim about payroll taxes that was true in 2012 could therefore be slightly off in 2013, which is the time-period question of the sort applied to the law itself rather than the data. The myth-buster’s discipline includes dating the statute, not just the statistics.

Supplemental Security Income provides the instructive contrast. Created in 1972, it pays benefits to aged, blind, and disabled people with little or no work history, and it is financed from general revenues, not from payroll levies. Its existence demonstrates that Congress knows how to build a program with individual eligibility determinations and means testing when it wants to; the retirement and disability programs’ choice of earnings related formulas without individual accounts is therefore a deliberate design decision, not an administrative accident. The two programs side by side show the full menu: general fund welfare on one hand, contributory social insurance on the other, and personal accounts nowhere on it.

Workers can verify the earnings record that determines their future benefit, which addresses the most practical version of the account confusion. The administration encourages filers to check their posted earnings for errors, since missing or misallocated wages reduce the computed benefit, and provides procedures to correct the record with pay stubs and employer documentation. What is being corrected is the history, not a balance: the remedy for an error is a higher computed benefit later, not a deposit into an account. Understanding that distinction is the difference between monitoring a number that matters and searching for a number that does not exist.

The trustees’ 2013 report projected the combined reserves depleting around 2033, after which incoming collections would cover about 77 percent of scheduled benefits, with the disability fund facing earlier depletion around 2016. The drivers were demographic and long visible: lower fertility, longer life expectancy, and the retirement of the large postwar birth cohorts. The projection assumed no legislative change, which history suggests is unlikely given the 1983 precedent. The relevant point for the myth is structural: a personal account system cannot become insolvent in this way, because each account holds its own balance; only a pay as you go system with legislated benefit promises can face a financing gap, which confirms what the system is.

The claim ledger

The table below collects the dozen claims examined above into a single reference. Each row states the claim as its holders would recognize it, gives the verdict, names the statute, decision, or dataset that decides it, and points to the series article carrying fuller treatment. The ledger is meant to be consulted the next time any of these contentions appears in argument: find the row, check the authority, and apply the four questions before replying.

Claim	Verdict of true, partly true, false, or genuinely unresolved	Specific statute, case or dataset that decides it	Series article carrying the full treatment
The Sixteenth Amendment was never properly ratified	False	U.S. Const. amend. XVI, ratification completed Feb. 3, 1913, proclaimed Feb. 25, 1913; Miller v. United States, 868 F.2d 236 (7th Cir. 1989)	This article
The amendment does not authorize a tax on wages	False	Brushaber v. Union Pacific R.R., 240 U.S. 1 (1916); 26 U.S.C. 61(a)(1); Commissioner v. Glenshaw Glass, 348 U.S. 426 (1955)	This article
Paying income tax is voluntary	False	Flora v. United States, 362 U.S. 145, 176 (1960); 26 U.S.C. 1, 6001, 6011, 6012, 6151; United States v. Tedder, 787 F.2d 540 (10th Cir. 1986)	This article
A higher bracket can reduce take-home pay	False	26 U.S.C. 1 rate schedule; marginal application to the dollars above each threshold	This article
The wealthy pay no tax	False	IRS Statistics of Income, tax year 2011: top 1 percent paid 35.1 percent of individual income tax	This article
The wealthy pay all the tax	Partly true	IRS Statistics of Income, tax year 2011; CBO distributional analysis (July 2012 edition current on the article date)	This article
Top income shares have risen sharply since the 1970s	Genuinely unresolved	Piketty-Saez tax return series and Sept. 2013 update vs. Auten-Gee-Turner (National Tax Journal, 2013); Saez-Zucman (2014/2016) vs. Auten-Splinter (2017) as later data points	This article
Corporations pay no tax	Partly true	Fiscal 2012 receipts, about 242 billion dollars; GAO-08-957 on zero liability years	This article
Workers vs. capital bear the corporate tax	Genuinely unresolved	CBO 75 percent capital and 25 percent labor; Treasury 82 and 18; JCT 75 and 25; Jennifer Gravelle (National Tax Journal, March 2013)	This article
The 1981 rate cuts paid for themselves	False	JCT staff estimate, S. Rept. 97-144: 688.1 billion dollar loss, FY1981-86, static; OMB and CBO historical tables	This article
The 2017 rate cuts paid for themselves	False	JCT JCX-67-17 (Dec. 18, 2017): 1.455 trillion dollar conventional loss, FY2018-27, as an explicitly dated later data point	This article
Payroll taxes accumulate in a personal account	False	42 U.S.C. 401(a) trust fund structure; Flemming v. Nestor, 363 U.S. 603 (1960)	This article

A practical checklist falls out of the method. When the next confident claim arrives, run the four questions before reacting: which levy, the income tax, the payroll levy, the corporate levy, or all federal levies combined; which income measure, adjusted gross, cash, or economic income; which unit, individual, tax unit, or household; which period, a single year or a lifetime. Then ask for the authority: the section number, the case name, or the dataset and table. Claims that cannot name an authority are not claims about the law; they are impressions. Claims that name one can be checked, and checking is a skill that improves with practice.

The symmetry of the corrections is worth restating as the article closes. The constitutional theories associated with skeptics of the system failed. The voluntariness theory associated with the same skeptics failed. The bracket fear, which afflicts no political camp in particular, failed. The slogans about the affluent, one from each direction, both overstated. The corporate claim split down the middle. The self financing claim failed against the scores. The personal account claim failed against the statute. No direction was spared and none was favored, because the record does not grade on affiliation. That evenhandedness is not neutrality for its own sake; it is what makes the verdicts trustworthy. A reader who can see a favored claim fail on the evidence is a reader equipped for the next dozen.

What the corrections add up to

A reader who entered holding any of these claims now holds something more useful than a verdict: a method. Ask which levy is under discussion before quoting a share. Ask which income measure the numbers use before declaring a trend. Ask which unit of analysis sorted the population before comparing groups. Ask which period the figures cover before drawing a conclusion about fairness. Those four questions do not favor either side of the underlying debates. They favor the reader, which is the point of the exercise.

Notice what the verdicts have in common across their different directions. The constitutional claims failed because the text, the certification, and a century of decisions say otherwise. The voluntariness claim failed because it confused an administrative method with a legal option. The bracket claim failed because of arithmetic. The distributional slogans failed because they each described one slice of a larger picture as if it were the whole. The corporate claim split, true in its narrow observation about particular companies in particular years, unresolved on the deeper question of incidence. The self financing claim failed against the published scores. The personal account claim failed against the statute’s trust fund structure. In every case the correction ran toward the record, not toward a side, and several of the corrections cut against positions this series’ own readers might hold. That symmetry is deliberate. A myths article that only debunks the other side’s myths is a pamphlet.

The unresolved labels deserve a final word, because they are the feature of this article most likely to be misunderstood. Labeling the incidence of the corporate levy unresolved, and labeling the measurement of top income shares unresolved, is not a failure to decide. It is the decision the evidence supports. On both questions, credentialed specialists using defensible methods reach different answers, the official scorekeepers publish their assumptions with caveats, and the honest summary names both sides and stops. Readers who want a single number for who bears the corporate levy, or a single series for top income shares, will have to choose among competing estimates with open eyes. Competence includes knowing when the record is silent, and the record is silent there.

The series thesis for this installment holds: correcting the record is part of making a reader competent. Competence does not mean holding approved opinions about rates or distribution. It means being able to hear a confident claim about American tax law, locate the statute or dataset that governs it, and grade it without assistance. The dozen claims above were chosen because they recur, because they mislead in both directions, and because each one surrenders to the same discipline of checking. Apply that discipline to the next dozen and the field becomes far less mysterious than its reputation suggests. Readers who keep a running notebook of claims tested and verdicts reached may find the VaultBook legislation study notebook a useful place to record them, alongside the statutes and decisions that decided each one.

Frequently Asked Questions

Q: Is it true the sixteenth amendment was never ratified under US tax law?

No. Ratification was completed on February 3, 1913, when the thirty-sixth of the forty-eight states then in the Union approved the amendment, and Secretary of State Philander C. Knox proclaimed it adopted on February 25, 1913. The discrepancies critics cite, differences in punctuation, capitalization, and wording between the congressional text and some state resolutions, are real as clerical facts but legally immaterial. The leading appellate decision, Miller v. United States, 868 F.2d 236 (7th Cir. 1989), examined those variations and upheld ratification, and the Tenth Circuit in United States v. Collins called the contrary argument “devoid of any arguable basis in law.” The Supreme Court built a century of income tax jurisprudence on the amendment beginning with the unanimous Brushaber decision of January 24, 1916. Raising the argument in filings can draw a 5,000 dollar frivolous submission penalty under section 6702, raised from 500 dollars in 2006.

Q: Is paying income tax voluntary under US tax law?

Not in the sense the claim means. The word voluntary in official usage describes self assessment, the administrative choice to have taxpayers compute their own liability rather than waiting for the government to assess them. The Supreme Court explained this in Flora v. United States, 362 U.S. 145, 176 (1960): “Our system of taxation is based upon voluntary assessment and payment, not upon distraint.” The duty itself is compulsory: section 1 imposes the tax, sections 6001 and 6011 require records and returns, section 6012 requires returns from earners above the threshold, and section 6151 requires payment when the return is due. Withholding under section 3402 collects much of the liability before any return is filed, and the Tenth Circuit confirmed in United States v. Tedder that Congress gave the Secretary power to enforce the laws “through involuntary collection.” Voluntary describes who does the arithmetic, not whether it must be done.

Q: Under US tax law, can a higher bracket lower your take home pay?

No. The rate schedule applies marginally, so each tier touches only the dollars inside it. Using the 2013 schedule for a single filer, the 25 percent tier began at 36,250 dollars of taxable income; a raise from 35,000 to 38,000 dollars left the first 36,250 dollars taxed exactly as before and applied 25 percent only to the 1,750 dollars above the line. The tax on the 3,000 dollar raise was 625 dollars, so the raise added 3,000 dollars before tax and 2,375 dollars after tax. The same arithmetic holds at every threshold, from the 10 percent bottom tier to the 39.6 percent top rate that applied above 400,000 dollars for single filers in 2013. The honest qualification is that benefit phaseouts elsewhere in the system can impose high effective burdens on additional earnings, but those are phaseout mechanics, not bracket mechanics. The bracket itself has never turned a raise into a pay cut under any schedule Congress has enacted.

Q: Do the rich pay no taxes under US tax law?

No. Published tabulations contradict the claim directly. The Internal Revenue Service Statistics of Income division reported that for tax year 2011, the top 1 percent of returns earned 18.7 percent of adjusted gross income and paid 35.1 percent of all individual income tax. The top 5 percent paid 56.5 percent of the income tax, the top 10 percent paid 68.3 percent, and the bottom half of returns combined paid 2.9 percent. Broader analyses that add payroll, corporate, and excise levies narrow the concentration but still show average federal rates rising with income, because the graduated income tax dominates the mix at the top while the payroll levy, which applies only to earnings below the annual wage base, flattens the picture. Individual cases vary widely around those averages: a high earner living on tax preferred investment income can face a lower effective percentage than a high earner living on wages. Those outliers are real but they describe individuals, not the distribution.

Q: Did Reagan’s tax cuts pay for themselves under US tax law?

The published official scores say no. Before passage of the Economic Recovery Tax Act of 1981, the staff estimate in Senate Report 97-144 projected a revenue loss of 688.1 billion dollars over fiscal years 1981 through 1986, and the estimate was explicitly static, meaning it assumed no behavioral feedback from the cuts. Fiscal history then showed receipts falling from 19.6 percent of gross domestic product in fiscal 1981 to 17.3 percent in fiscal 1984, while deficits widened past 5 percent of output in the mid 1980s. Professional estimates from the Congressional Budget Office, the Joint Committee on Taxation, and the Treasury Department consistently found that added economic activity offset only a fraction of the static loss. Feedback effects are real and were included in the estimates, so the claim fails on the scorekeepers’ own terms rather than on terms its advocates were never offered.

Q: Do corporations pay no tax under US tax law?

Not in the aggregate. The corporate income levy produced about 242 billion dollars in fiscal year 2012, roughly 10 percent of federal receipts, and no fiscal year on record shows the business sector escaping the levy entirely. The narrower observation behind the claim is real: a 2008 Government Accountability Office study, GAO-08-957, found that about 55 percent of large corporations under American control reported no federal tax liability in at least one year between 1998 and 2005, reflecting losses, credits, and timing provisions the Code itself provides. A zero in one year for one company is not a zero for the sector across time. The deeper question of who ultimately bears the levy, shareholders through lower returns or workers through lower wages, is genuinely unresolved. The Congressional Budget Office allocates 75 percent to capital and 25 percent to labor, the Treasury uses 82 and 18, the Joint Committee on Taxation uses 75 and 25, and the academic literature spans the full range.

Q: Are wages income under US tax law?

Yes. Section 61(a)(1) of the Internal Revenue Code expressly includes compensation for services in gross income, and the courts have read the term broadly for a century. In Eisner v. Macomber, 252 U.S. 189 (1920), the Supreme Court defined income as gain derived from capital, from labor, or from both combined. In Commissioner v. Glenshaw Glass, 348 U.S. 426 (1955), the Court adopted the standard test of undeniable accessions to wealth, clearly realized, over which the taxpayer has complete dominion, which wages satisfy without strain. In Lucas v. Earl, 281 U.S. 111 (1930), the Court held that earnings from personal services are taxed to the person who earns them. The argument that wages fall outside income has been raised many times and has lost every time, with courts describing it as frivolous and imposing sanctions, including the 5,000 dollar penalty under section 6702, on those who press it.

Q: Is it a myth that the IRS can seize your home under US tax law?

Partly. The collection powers are real and specific. Section 6321 creates a federal tax lien on all property and rights to property when tax goes unpaid after demand. Section 6331 authorizes the Secretary, after ten days’ notice and demand, to collect by levy, which the statute defines to include the power of distraint and seizure by any means, reaching property whether real or personal. But a principal residence carries a special statutory protection: under section 6334(e)(1)(A), the agency may not levy on a principal residence without the written approval of a judge or magistrate of a United States district court. In practice, seizures of primary residences are extraordinarily rare and occur only after extensive notice, opportunity to pay or arrange alternatives, and exhaustion of less drastic collection steps. So the folk version of the claim, that the agency routinely takes homes over ordinary balances, is false, while the narrow legal proposition that the power exists under tight judicial restriction is true.

Q: Is it true that filing an extension gives you more time to pay under US tax law?

No. Section 6081 permits an automatic extension of time to file the return, commonly six months, but the extension applies only to the paperwork. Section 6151 keeps the payment deadline where the statute puts it, generally April 15 for individuals, and interest accrues on any unpaid balance from that date regardless of the extension. A late payment addition under section 6651 also applies, though it is reduced for taxpayers who paid most of the liability on time. The extension is therefore useful for filers who need more time to complete an accurate return, but it is not a deferral of the underlying obligation. Anyone expecting to owe should remit an estimated payment with the extension request to stop interest from running. The folk belief treats the filing deadline and the payment deadline as one date; the Code treats them as two.

Q: Is it true that a tax refund means you paid no tax under US tax law?

No. A refund means withholding or estimated payments exceeded the final liability, not that the liability was zero. Employers withhold under section 3402 based on tables and the employee’s withholding certificate, which are approximations, and section 31 credits those withheld amounts against the tax computed on the return. When the credit exceeds the liability, the difference comes back as a refund; when it falls short, the filer pays the balance. Either way the underlying tax was computed the same way. The confusion treats the settlement of the account as the account itself. Large refunds usually indicate over withholding during the year, which is an interest free loan to the government rather than evidence of a light burden, and large balances due usually indicate under withholding rather than a heavy one.

Q: Is it true that most workers pay the alternative minimum tax under US tax law?

No. The alternative minimum tax, imposed by sections 55 through 59, was designed as a backstop to keep high income filers with large preferences from eliminating liability, and it reaches only a small fraction of returns. For 2012, roughly 4 million taxpayers paid it, out of more than 140 million individual returns filed. The tax bites hardest at upper middle incomes in high tax states, because state and local tax deductions are disallowed in its computation, and Congress enacted a series of annual patches to keep the exemption amounts from sweeping in millions more as nominal incomes rose. The American Taxpayer Relief Act of 2012 then indexed the exemption permanently. Most households never compute it at all. Its reputation as a mass tax comes from projections of what would happen without those patches, not from who actually pays it.

Q: Is it true that the estate tax reaches most inheritances under US tax law?

No. The estate tax under section 2001 applies only to estates above a large exemption, 5.25 million dollars per decedent in 2013, with a top percentage of 40 percent on amounts above that line. Because of the exemption, only a tiny fraction of decedents, roughly one or two in a thousand, leave estates large enough to owe anything, and married couples can combine exemptions to shield more than 10 million dollars. Most inherited wealth therefore passes with no federal estate levy at all, though heirs should note that the income tax still applies to certain inherited items such as retirement account distributions. The tax’s reputation as a levy on ordinary inheritances comes from its political salience, not its reach. As a revenue source it is small, supplying about 1 percent of federal receipts.

Q: Is it true that unreported cash income escapes taxation under US tax law?

No. Section 61 defines gross income as all income from whatever source derived, and the Supreme Court has applied that language to unlawful as well as lawful receipts, famously in James v. United States, 366 U.S. 213 (1961), concerning embezzled funds. Failing to report cash earnings is not a gap in the law but a violation of it: section 7201 makes willful evasion a felony, and civil fraud penalties under section 6663 add 75 percent of the underpayment. Prosecutors routinely build cases from bank records, lifestyle evidence, and the specific items method when books are absent. The folk belief that cash leaves no trail confuses difficulty of detection with absence of liability. Detection is harder, which is why enforcement resources target it, but the legal duty attaches the moment the income is received.

Q: Is it true that charitable giving always reduces your tax under US tax law?

No. Section 170 permits a deduction for contributions to qualifying organizations, but the deduction helps only filers who itemize. Taxpayers who claim the standard deduction, 6,100 dollars for single filers and 12,200 dollars for joint filers in 2013, receive no federal benefit from their giving, and most filers take the standard amount. Even for itemizers, limits apply: cash gifts to public charities are generally capped at 50 percent of adjusted gross income, with lower caps for certain property gifts and private foundations, and excess amounts carry forward. Noncash gifts above 5,000 dollars require a qualified appraisal. The deduction also has no value against the alternative minimum tax computation for some gifts. Giving remains admirable, but its tax effect is conditional on itemizing, on the caps, and on the rest of the return.

Q: Is it true that bartering services avoids income tax under US tax law?

No. The Internal Revenue Service ruled in Revenue Ruling 79-24 that the fair market value of services received in a barter exchange is includible in gross income under section 61, and courts have sustained the position. A plumber who fixes a dentist’s pipes in exchange for dental work has income equal to the value of the dental work, and the dentist has income equal to the value of the plumbing. Barter exchanges are required to report transactions on information returns, and members receive statements of their trade credits. The medium of exchange does not change the tax result: income realized in kind is income all the same. Informal neighborhood swaps rarely draw attention, but the legal treatment is settled, and larger barter activity conducted through exchanges is reported systematically.

Q: Is it true that state income taxes follow the federal definition under US tax law?

No, though most do as a practical matter. Each state writes its own tax code, and none is required to follow the federal definition of income. Most states begin their computation with federal adjusted gross income or federal taxable income and then apply state-specific additions and subtractions, because piggybacking on the federal base is administratively convenient. States differ in how they conform: some adopt the Code as amended on a rolling basis, automatically picking up federal changes, while others conform to the Code as of a fixed date and must legislate to catch up. A few states use entirely independent definitions. The result is that a provision’s federal treatment does not dictate its state treatment, and multi-state filers can face genuinely different liabilities on the same income.

Q: Is it true that Social Security benefits are tax free under US tax law?

No. Section 86, added by the 1983 amendments and expanded in 1993, makes up to 85 percent of benefits includible in gross income for beneficiaries above specified income thresholds. Below those thresholds benefits remain untaxed, so lower income retirees generally owe nothing on them, while higher income beneficiaries include a portion. The thresholds are not indexed for inflation, which means more beneficiaries cross them over time as nominal incomes rise. The provision was enacted as part of the 1983 financing repair and extended a decade later, and it remains one of the less understood features of retirement taxation. Anyone estimating retirement cash flow should account for it rather than assuming the benefits arrive untouched, and the revenue from the provision flows back to the trust funds.

Q: Is it true that businesses may deduct every expense under US tax law?

No. Section 162 permits deductions only for ordinary and necessary expenses paid or incurred in carrying on a trade or business, and the Code layers specific disallowances on top of that standard. Fines and penalties paid to a government are nondeductible under section 162(f). Bribes and kickbacks are barred. Personal, living, and family expenses are excluded by section 262 even when a sole proprietor incurs them. Capital expenditures must generally be recovered over time through depreciation rather than deducted at once. Interest deductions carry their own limits. The ordinary and necessary test itself excludes extravagant spending. Planning that assumes every outlay reduces taxable income will overstate the benefit and invite adjustment on examination.

Q: Is it true that the payroll tax stops at the wage cap under US tax law?

It depends on which portion. The Social Security portion of the Federal Insurance Contributions Act levy applies only to earnings up to the annual wage base, 113,700 dollars in 2013, with the base adjusted yearly for wage growth. Earnings above that line face no Social Security levy. The Medicare hospital insurance portion has no cap: 1993 legislation removed its wage base, so every dollar of wages faces the Medicare percentage. The combined employee percentage in 2013 was 7.65 percent up to the base and 2.35 percent above it, with employers matching. Self employed workers pay both halves under the Self Employment Contributions Act. The cap is the feature that makes the payroll levy regressive at the top of the earnings distribution, since investment income escapes it entirely.

Q: Is it true that the IRS has only three years to assess more tax under US tax law?

Mostly, but the exceptions matter. The general rule under section 6501 is three years from the date the return was filed or due, whichever is later. If the filer substantially understates gross income by more than 25 percent, the period extends to six years. There is no time limit at all for a fraudulent return or for a year in which no return was filed, and the period can be extended by agreement between the taxpayer and the Service. Collection after assessment has its own ten-year clock under section 6502. These limits are why record-keeping guidance commonly suggests retaining returns and supporting documents for at least three years, and longer when the exceptions might apply. The statute of limitations does not forgive the tax; it bars the government from assessing it after the window closes.