The Government Once Prosecuted Men Under a Rule That No Longer Existed

Suppose you were charged with breaking a law, and halfway through the case the government quietly discovered that the law had been repealed. By accident. Months earlier. And that nobody in the building had noticed because nobody in the building could find a complete copy of the rules.

That is not a satire. It is, in outline, what happened in the winter of 1934 and 1935, in a case about oil. It is one of the more astonishing episodes in the history of American administration, and it produced something we all take for granted: a daily newspaper that exists to publish the government's own rules. To understand why that newspaper had to be invented, we have to go back to the first year of the New Deal, when Washington was making law faster than it could keep track of it.

The problem: law made by order, filed nowhere

Start with the premise most Americans hold without ever stating it: if the government can punish you for something, you can look up what the something is. In 1933 that premise was not reliably true.

The National Industrial Recovery Act, passed in June 1933, authorized the President to approve codes of fair competition for industries, and to issue executive orders, regulations and administrative rulings to carry out the act. The stories of the codes are told in A Brooklyn Poultry Firm Beat the New Deal Nine to Nothing in 1935. What matters here is the volume. Within months there was a flood of executive orders, administrative orders and interpretations that modified them. Many were issued as typed or mimeographed sheets. Some were distributed to trade associations and not to anyone else. Some existed only in the files of the agency that issued them.

There was no single place where a lawyer, a business owner or a prosecutor could go to find which rules were in force on a given day. Rules were amended by later rules, and the amendments might themselves be amended. A lawyer could spend a day looking for a regulation that his client was accused of breaking and come away unsure whether it still existed.

The oil: what a "hot oil" rule was

The oil industry gives the story its center. In the early 1930s the discovery of the huge East Texas field had flooded the market with crude oil, and prices had fallen to almost nothing. The oil-producing states, Texas and Oklahoma among them, set quotas on how much a producer could pump. Producers who exceeded the state quota produced what was called "hot oil," oil that had been pumped in violation of state law.

Section 9(c) of the National Industrial Recovery Act authorized the President to prohibit the transportation in interstate commerce of oil produced in excess of the amount allowed under state law. The President did so by an executive order and by a code of fair competition for the petroleum industry, with penalties for violations. The federal government then began prosecuting producers and shippers who moved hot oil across state lines. The defendants in the Supreme Court case were two Texas oil companies.

The accident

Now the embarrassment. The prohibition on shipping hot oil had been part of the petroleum regulations issued in 1933, but the specific provision the defendants were charged under was no longer in the regulations when they were charged. According to the National Archives' own history of the Federal Register, the two Texas oil companies in the case had been charged with violating a provision that technically did not exist at the time, and at the oral argument on December 10, 1934, the Justices focused on that defect. The next day the Harvard Law Review published an article by Erwin Griswold, "Government in Ignorance of the Law," which pleaded for better publication of executive legislation. Retellings of exactly who found the missing provision and when vary, so this post does not rely on them. What is not in dispute is the substance: the government was enforcing a rule that could not be found in the regulations then in force.

That is the vivid part, and it is the part that gave the Federal Register its political push. But the case itself is more important than the anecdote.

Panama Refining Co. v. Ryan: the first nondelegation holding

The case was Panama Refining Co. v. Ryan, and the Supreme Court decided it on January 7, 1935. Chief Justice Hughes wrote the opinion for an 8-1 majority, with Justice Cardozo in dissent. The question the Court actually decided was whether Section 9(c) was a valid delegation of legislative power.

The majority said it was not. The statute told the President he could prohibit the interstate shipment of hot oil, and that was all. It did not say under what circumstances he should do it, or what findings he had to make, or what policy should guide him. It declared no standard. Hughes put the problem this way: Congress had left the matter to the President's unfettered discretion, with no policy or criterion to guide him. The Court observed that Congress had not required the President to find any fact before acting, or to adopt a rule of conduct. In effect, Congress had said, "Prohibit hot oil if you think it wise."

This was the first time in American history that the Court struck down a federal statute on the ground that it unconstitutionally delegated legislative power. It was followed less than five months later by Schechter, where the justices struck down the code-making system in its entirety, and the pair together are the only nondelegation victories the Court has ever handed down. The doctrine's background, including the Constitution's text and the later "intelligible principle" test, is explained in The Nondelegation Doctrine Says Congress Cannot Hand Its Job to Someone Else.

Justice Cardozo's dissent is worth a moment, because he was no enemy of limits on government. He thought the statute, read in context, did supply guidance, and that the Court was being too demanding. Five months later he joined the unanimous Schechter decision, once the delegation he faced was vastly broader. The difference between the two cases tells you the line is one of degree, and that even the Court's most sympathetic justices thought there was a line.

The cure: the Federal Register Act of 1935

Congress did not wait for the courts to finish the lesson. The hot oil episode showed that the government's own lawyers could not find its own rules, and that a citizen charged with a violation might never have had a way to learn what the rule said.

The Federal Register Act was passed in 1935, signed on July 26, 1935. It created a system for publishing, in a uniform way, the executive orders, proclamations and administrative regulations that had general applicability and legal effect. The key idea was publication as a condition of validity: under the act, a covered document could not be valid against any person until it was filed with the Federal Register. This is the principle of notice, and it is one of the oldest rights in English and American law. The people cannot obey what they cannot read.

The first issue of the Federal Register appeared on March 14, 1936. It has appeared on business days ever since. The act also arranged for the permanent codification of regulations by subject, and in 1938 the President authorized a new publication, the Code of Federal Regulations. The CFR is the set of books, and now an online database, in which the regulations in force are organized by topic into titles. The explainer on Your Questions About the Code of Federal Regulations, Answered Plainly shows how to read it.

Credit where it is due. Few reforms of the New Deal era are as plainly good as this one. Nobody can seriously defend the opposite proposition, that citizens should be punished under rules they cannot find. Conservatives can and should say so, and then ask what it reveals.

What the Register tells us about government now

The Federal Register was built to solve a problem of 1935, which was that rules could not be found. Today it solves that problem, and it also measures another. Its daily volume is the best single gauge we have of how much law the executive branch produces.

Take the numbers. The Federal Register published 3,248 final rules in 2024 and 2,441 final rules in 2025. Final rules are the finished regulations, after proposal and comment, that carry the force of law. Here is the arithmetic. The difference between the two years is 3,248 minus 2,441, which is 807 rules. Divide 807 by 3,248 and you get about 0.248, so the count fell by roughly a quarter from one year to the next.

Now spread the 2,441 over the year. The Register is published on business days, and there are roughly 250 of those in a typical year, so as a rough illustration, 2,441 divided by 250 is a little under ten final rules on each publication day. For 2024, 3,248 divided by 250 is about thirteen. Those are averages for illustration, since the days are not evenly loaded, and they count only the final rules, not the proposed rules, notices or presidential documents that fill the rest of each issue.

Compare that with the other side of the ledger. Congress enacts, in an ordinary year, a few hundred public laws, a figure you can check at Congress.gov, and a good many of them are minor, such as the naming of post offices. The point of the comparison is not the exact ratio. It is that the rules outnumber the statutes by a wide margin, and the people who write the rules were not elected. The site's own red tape page tracks the flow, and Agencies Write Thousands of Rules a Year While Congress Passes Hundreds of Laws sets the two columns next to each other.

A defender of the system will say that most final rules are routine: they adjust a deadline, update a fee table or set a seasonal fishing limit. That is true, and it is a fair point. But the routine rules and the major rules come out of the same machine, and the machine's output is too large for any member of Congress, any citizen or any single lawyer to read. In 1935 the government could not find its rules. In our time the government can find them, and the citizen cannot read them all. Publication solved the first problem and not the second.

The steelman: is volume really a problem?

The strongest form of the objection is this. One: a large economy needs a large body of rules, and the number of rules is not a measure of their burden. Two: many rules deregulate, simplify or grant exceptions. Three: the Register's growth in pages reflects the Register's thoroughness, which is a virtue. Four: the alternative to agency rules is not fewer rules but less specific statutes, which produce more litigation.

Each of those points contains truth. A raw count is a crude instrument, and I would not hang an argument on the count alone. Some of the most important rules are one-page rules, and some of the longest are technical. But the reasoning does not answer the question the Constitution asks, which is not "how many?" but "who decided?" Whether the number is 2,441 or 3,248, the people who wrote those rules answer to the President through a long chain, and to Congress through a thinner one. And the rule you must obey is the one you cannot easily find, in a database you cannot easily search, on a subject you cannot easily follow. The article on The Small White House Office That Reads Every Major Rule Before You Do describes the one office that does read the major ones.

What the oil case teaches, point by point

1. Notice is a precondition of justice. The rule of law means that the law is knowable. The hot oil case showed a government unable to meet that minimum.
2. Delegation without standards invites chaos. Congress had written one broad sentence and left the rest to the executive. The consequence was thousands of pages of orders, and the prosecutors lost track of them.
3. Courts can enforce the structure. The Supreme Court did not wait for Congress to fix the problem. It held, in January 1935, that the statute went too far.
4. Reform follows embarrassment. The Federal Register exists because an agency was caught out. The lesson is that sunlight needs a scandal, and scandals need people paying attention.
5. Transparency is necessary and not sufficient. Publishing rules makes them findable. It does not make them accountable.

The Founders' standard

The Declaration of Independence lists, among its grievances against the King, that he had erected "a multitude of New Offices, and sent hither swarms of Officers to harass our People, and eat out their substance." Every American can recognize the pattern. The Founders were not complaining that there were officers. They were complaining that the officers made demands the people could not answer and could not escape.

A republic runs on law that its citizens can read, and that their elected representatives wrote. Madison put the point in Federalist 62, observing that a law so voluminous that it cannot be read, or so incoherent that it cannot be understood, cannot be a rule of action, because what cannot be known cannot be a rule. The Register fixed the second defect. We are still working on the first. The post on Federalist 62 Warned That Mutable Laws Betray a Free People reads that essay line by line.

The case in brief

1. In Panama Refining Co. v. Ryan, decided January 7, 1935, the Court held that the NIRA's hot oil provision gave the President power with no standard, the first nondelegation holding in the Court's history.
2. The litigation exposed that the provision being enforced was missing from the regulations then in force, because no one could reliably find the rules.
3. Congress responded with the Federal Register Act of July 26, 1935, followed by the first Federal Register issue on March 14, 1936 and the authorization of the Code of Federal Regulations in 1938.
4. Publication solved the problem of finding rules. It did not solve the problem of volume: the Register recorded 3,248 final rules in 2024 and 2,441 in 2025, written by people the voters cannot remove.
5. The remedy is the one the Founders pointed to: major rules should be decided by the legislature, and the rest should stay within standards the legislature actually set.

What you can do: open the Federal Register on the site of the National Archives, find one proposed rule in a field you care about, and file a public comment before its deadline, then check back to see whether the final rule answered you.