Federalist 47 Defines Tyranny as All Powers Gathered Into the Same Hands

Imagine a basketball game in which one official writes the rulebook the morning of the game, blows the whistle on your fouls during it, and then hears your appeal afterward in a room with no one else present. You might win that game. It would be a miracle. And if you lost, you would not need a law degree to know what had gone wrong.

James Madison would have recognized the problem at once. He gave it a name in Federalist 47, published in early 1788, and the name was not "inefficiency" or "bad management." The name was tyranny. Not tyranny as a figure of speech, the way people say their homeowners' association is a tyranny, but tyranny as a definition, precise enough to test any government against. The definition does not care whether the officials holding the power are kings, elected legislators or civil servants with excellent credentials. It cares only about where the power sits.

That makes Federalist 47 the most useful paper in the series for understanding the modern administrative state. Let us read it closely and then apply its test.

The charge Madison was answering

The Anti-Federalists made a serious accusation against the proposed Constitution: that it violated the separation of powers. The Senate shared in appointments and treaties, which are executive acts. The president had a veto, which is a legislative act. The Senate tried impeachments, which is a judicial act. The powers were mingled, they said, and mingled powers were the road to despotism.

Madison does not dismiss the principle behind the charge. He embraces it, in words that are as strong as anything in American political writing:

No political truth is certainly of greater intrinsic value, or is stamped with the authority of more enlightened patrons of liberty, than that on which the objection is founded. The accumulation of all powers, legislative, executive, and judiciary, in the same hands, whether of one, a few, or many, and whether hereditary, self-appointed, or elective, may justly be pronounced the very definition of tyranny.

Look at the qualifications he piles on. "Whether of one, a few, or many": tyranny is not just a monarch; a committee or a mass assembly can be a tyrant. "Whether hereditary, self-appointed, or elective": being elected does not cure it. A tyranny chosen by a majority is still a tyranny. This is the point modern discussions most often miss. The question is not whether the officials are accountable to voters in some distant way. The question is whether the three kinds of power are held by the same hands.

Madison then says that if the Constitution really did accumulate the powers in this way, "no further arguments would be necessary to inspire a universal reprobation of the system." He is confident it does not. So the paper's task is to show what the principle actually requires.

Montesquieu, read correctly

Madison's source is the French philosopher whose The Spirit of the Laws, published in 1748, gave the separation of powers its classic modern form:

The oracle who is always consulted and cited on this subject is the celebrated Montesquieu.

Madison quotes Montesquieu's warnings as they were rendered into English at the time:

When the legislative and executive powers are united in the same person or body, there can be no liberty, because apprehensions may arise lest THE SAME monarch or senate should ENACT tyrannical laws, to EXECUTE them in a tyrannical manner.

Were the power of judging joined with the legislative, the life and liberty of the subject would be exposed to arbitrary control, for THE JUDGE would then be THE LEGISLATOR. Were it joined to the executive power, THE JUDGE might behave with all the violence of AN OPPRESSOR.

The capital letters are in the original paper. Madison wanted readers to see the mechanism. A legislator who also executes can write laws designed for the way he means to enforce them. A judge who is also a legislator decides cases under rules he wrote. A judge who is also the executive can punish whomever he has already decided to prosecute.

Then comes the key interpretive move. Montesquieu's model was the British constitution, and in Britain the powers were not perfectly separate. The king was part of the legislature through his assent to bills. He appointed the judges, who could be removed on the address of both houses of Parliament. The House of Lords was the highest court of appeal. So Montesquieu could not have meant that the branches must have nothing to do with each other. Madison draws the conclusion:

his meaning, as his own words import, and still more conclusively as illustrated by the example in his eye, can amount to no more than this, that where the WHOLE power of one department is exercised by the same hands which possess the WHOLE power of another department, the fundamental principles of a free constitution are subverted.

That is the test. Not any overlap, but the whole power of one department held by the hands that hold the whole power of another.

The state constitutions as evidence

Madison then goes state by state through the constitutions Americans had written since 1776. Every one of them declared the separation of powers in some form. Not one kept the powers absolutely separate.

The Massachusetts constitution of 1780, drafted largely by John Adams, contains what may be the most emphatic statement of the principle ever written into an American charter:

In the government of this commonwealth, the legislative department shall never exercise the executive and judicial powers, or either of them: the executive shall never exercise the legislative and judicial powers, or either of them: the judicial shall never exercise the legislative and executive powers, or either of them: to the end it may be a government of laws and not of men.

Yet, Madison observes, even Massachusetts gave its governor a qualified veto, had its senate sit as a court of impeachment, and had its judges appointed by the executive. New Hampshire's constitution, he notes, explicitly qualified the principle by saying the departments should be kept as separate as the nature of a free government would admit, consistent with the chain of connection that binds the constitution together. The other states mixed their powers in various ways, some far more than the proposed federal Constitution did.

The lesson Madison draws is that the Americans who loved the separation of powers most had never understood it to mean airtight compartments. They understood it to mean that no one set of hands should hold the whole of two powers. The next paper, Federalist 48, explains why some mixture is not merely permitted but necessary: parchment barriers alone cannot hold the branches apart. Each needs enough of a share in the others' work to defend itself. Federalist 51 completes the thought with ambition counteracting ambition. The story of how Adams built his constitution is at John Adams Wrote the Oldest Working Written Constitution in the World.

Applying the test to the administrative state

Now apply Madison's test to an agency. Here is a hypothetical, labelled as such, built on the ordinary structure of federal regulation.

Picture a federal commission created by statute with a broad mandate, say, to prevent "unfair practices" in an industry.

1. It writes the rules. Under the Administrative Procedure Act of 1946, the commission publishes a proposed rule, takes comments and issues a final rule. That rule binds the public with the force of law. Violating it carries penalties. That is legislative power in everything but name.
2. It enforces the rules. The commission's staff investigates, issues subpoenas and decides whom to charge. That is executive power.
3. It judges the violations. The case can be tried before an administrative law judge employed by the agency, and the commission itself may review the decision on appeal. That is judicial power.
4. Until 2024, it also interpreted its own statute with deference. Under the Chevron doctrine, when the statute was ambiguous, reviewing courts deferred to the agency's reasonable reading of it.

Count the hands. The same commission makes the law, prosecutes under it and adjudicates the case, and then, under Chevron, received deference from the courts on the meaning of the law that defined its own power. That is the basketball referee from the opening, with a government seal.

Madison's test is not whether there is any overlap. It is whether the whole of one power and the whole of another are in the same hands. For the matters within its jurisdiction, a commission like this holds the whole lawmaking power that Congress delegated, the whole enforcement power, and the first and often final adjudication. The scale is not trivial. The Federal Register published 3,248 final rules in 2024 and 2,441 in 2025. You can follow new ones on the site's red tape page.

The courts begin to apply Madison

The Supreme Court has started to take the problem seriously.

• Loper Bright Enterprises v. Raimondo (2024) overruled Chevron. Courts now decide for themselves what a statute means, rather than deferring to the agency's reading. That restores one judicial function to the judiciary. See Chevron Began as a Win for Reagan's EPA, Which Makes Its Fall So Ironic.
• SEC v. Jarkesy (2024) held that when the Securities and Exchange Commission seeks civil penalties for securities fraud, the Seventh Amendment entitles the defendant to a jury trial, which means the case belongs in a court, not before the agency's in-house judge. That restores part of the judicial power to the judiciary and the jury. See SEC v. Jarkesy Put the Jury Back Between Citizens and Agency Courts.
• West Virginia v. EPA (2022) applied the major questions doctrine: an agency may not claim power over questions of vast economic and political significance without clear authorization from Congress. That pushes lawmaking back toward the legislature.
• Seila Law v. Consumer Financial Protection Bureau (2020) held that the president must be able to remove the single director of the CFPB at will, so an agency wielding significant executive power cannot be insulated from the elected executive. That limits the reach of Humphrey's Executor v. United States (1935), the case that had allowed for-cause protection for members of multi-member commissions.

None of these cases abolishes the administrative state. Each one pulls a piece of power back toward the branch the Constitution assigned it to. The older roots of the nondelegation principle are in The Sick Chicken Case of 1935 and the Doctrine That Could Rein In Agencies.

The objection: Madison himself allowed the blending

The strongest objection, in terms its defenders would accept, is this. Federalist 47 does not condemn mixed powers; it defends them. Madison's whole point was that the Constitution's blending of powers was acceptable because no department held the whole of another's power. Agencies fit that description. Congress can rewrite or repeal their statutes. The president appoints and, in most cases, removes their heads. Courts review their rules and decisions. The Congressional Review Act of 1996 gives Congress a fast-track procedure to overturn new rules. And a modern economy requires expert bodies that can write detailed technical rules no legislature has time to draft. Madison's test, properly applied, approves of the administrative state.

That is a real argument. Here is the answer.

1. The test is about what the hands hold, not what can eventually check them. Madison asked whether the same hands hold the whole of two powers. That Congress could someday repeal a statute does not change what the agency holds today. A king who could be deposed by Parliament was still a king while he reigned.

1. The checks are weaker than they look. The Congressional Review Act requires a joint resolution that the president can veto, so in practice it is used mainly when a new president's party wants to overturn the previous administration's late rules. Congress rarely rewrites an agency's organic statute. And until Loper Bright, judicial review of statutory meaning was itself deferential.

1. The Constitution's blending was deliberate and listed; the agency's is general. The president's veto, the Senate's role in treaties and appointments and its trial of impeachments are specific, enumerated exceptions written into the text, each designed to help one branch defend itself. An agency's combination of rulemaking, enforcement and adjudication is not a specific exception. It is a general fusion across an entire subject area.

1. Expertise is an argument for advice, not for power. Congress can hire experts, hold hearings, and ask agencies to propose detailed rules that Congress then enacts. The REINS Act idea, requiring congressional approval of major rules, is one version. Expertise does not require that the expert also be the legislator and the judge.

1. Madison's own qualifications cut against the agencies. He wrote "whether hereditary, self-appointed, or elective." Being supervised at a distance by elected officials does not make an accumulation of powers acceptable. Being elected does not even do that.

The verdict: Madison's test permits the Constitution's specific, defensive blends. It condemns general fusions of whole powers in one body. Much of the administrative state falls on the wrong side of that line, and the courts are beginning to say so. A related essay on the same theme is Madison Called It the Very Definition of Tyranny, and Agencies Hold All Three.

Why it matters to a family

This is not a lawyers' parlor game. Picture a family business, hypothetically, that receives a notice from an agency charging a violation of a rule the agency wrote. The family's case will be heard first by a judge the agency employs. The appeal goes to the commissioners who authorized the charge. By the time a real court sees it, the family has spent its savings on lawyers, and the court will review the agency's findings of fact with deference. Every step of that process may be legal. Madison would still recognize it.

Separation of powers is the reason a citizen accused by the government gets a hearing before someone who is not the accuser and under a law that the accuser did not write. It is a protection for ordinary people far more than for the powerful, who can afford to wait out the process. The companion argument on how Congress itself contributes to the problem, by passing vague laws and leaving the details to agencies, is in Federalist 62 Warned That Mutable Laws Betray a Free People. And the case for one accountable executive at the top of the enforcement machine is in Federalist 70 Read Closely, Hamilton's Case for Energy in One Executive.

The case in brief

1. Federalist 47 defines tyranny as the accumulation of legislative, executive and judicial power in the same hands, however those hands are chosen.
2. Madison read Montesquieu to forbid not every overlap but the holding of the whole power of one department by those who hold the whole power of another.
3. Every state constitution of the period declared the principle, and none kept the powers airtight; the blends they allowed were specific and defensive.
4. A federal agency that writes binding rules, enforces them and judges violations fuses whole powers across its subject area, which is the accumulation Madison condemned.
5. Loper Bright, Jarkesy, West Virginia v. EPA and Seila Law have begun to return those powers to the branches the Constitution assigned them to.

What you can do: read one proposed rule on the site's red tape page, file a public comment on it before the deadline, and send a copy to your representative asking whether Congress should vote on it.