Federalism as experiment.

In 1932, Justice Louis Brandeis wrote a dissent in a case about ice.

The case was *New State Ice Co. v. Liebmann*, and it concerned whether Oklahoma could require a license to sell ice. An existing ice company wanted to block a competitor from entering the market. Oklahoma said the state could regulate entry into the ice trade as a public utility. The Supreme Court, 6-2, said it could not. The regulation violated the due process clause of the Fourteenth Amendment.

Brandeis thought the Court was wrong. His dissent is remembered not for its resolution of that narrow question, but for a single paragraph buried near the end: "It is one of the happy incidents of the federal system that a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country."

The phrase "laboratories of democracy" has since become one of the most cited and least understood ideas in American constitutional law.

What Brandeis actually meant

Brandeis was not making an ideological argument for decentralization. He was making an epistemological argument for institutional humility.

The context matters. The case was decided in 1932 — three years into the Great Depression, two years before the New Deal would reshape American regulatory law. The Court majority believed that state regulation of private business was presumptively unconstitutional. Brandeis disagreed, and the laboratory metaphor was his counter-argument: the Constitution should leave states room to try different approaches, because no one — not the Court, not Congress, not economists — yet knew which approaches to the Depression would work.

The laboratory, in Brandeis's metaphor, is not a libertarian sanctuary. It is a space for experimentation under conditions of genuine uncertainty. The argument is not "states should be free from federal interference." The argument is "when we do not know the answer, we should be able to try different things and find out."

The Brandeis paradox

Here is what makes Brandeis complicated as a patron of decentralization: he was simultaneously one of the strongest advocates for federal and state regulation of private economic power in the early twentieth century. He supported antitrust law. He supported consumer protection. He believed that concentrated private power — in banks, in corporations, in trusts — was as much a threat to self-government as concentrated public power.

He was not opposed to regulation. He was opposed to premature certainty about what regulation should look like.

This is why invoking Brandeis to oppose all federal intervention misreads him. And why invoking him to justify any state regulation, however designed, also misreads him. The laboratories metaphor is an argument for trying things, measuring them, and retaining the option to change — not an argument for any particular answer.

The Tenth Amendment and what it does

The Tenth Amendment reads: "The powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people."

It is the most quoted and least operative amendment in the Bill of Rights. For most of American constitutional history, it has functioned less as an enforceable rule and more as a statement of constitutional structure — a reminder that the federal government is one of limited, enumerated powers, not the default sovereign.

The Rehnquist Court in the 1990s gave the Tenth Amendment more operative force than it had held in decades. *United States v. Lopez* (1995) and *United States v. Morrison* (2000) held that Congress had exceeded its commerce power in regulating guns near schools and violence against women, respectively. *Printz v. United States* (1997) held that the federal government could not commandeer state executive officers to enforce federal law.

These decisions are controversial and contested. What they established, as a constitutional matter, is that states retain a meaningful domain of exclusive authority — that the federal government cannot simply instruct states to implement federal programs using state personnel and state resources.

For the purpose of institutional experimentation, this matters less as a legal doctrine and more as a structural fact: American federalism, even in its current form, leaves substantial room for state and local variation. States can enact different tax structures, different land use rules, different approaches to public health, education, criminal justice, and social services. The range of permissible variation is large.

Decentralization as ideology vs. decentralization as learning architecture

The political use of federalism arguments tends to be instrumental. Conservatives invoke federalism when federal policy moves left; progressives invoke it when states move right. The invocation tracks the speaker's preferred outcome, not a consistent theory of governance.

There is a different way to think about decentralization — not as a constitutional boundary to be defended or attacked, but as a structure for learning.

From this perspective, the relevant question is not "how much power should the states have?" The relevant question is: "what institutional structure allows us to discover, over time, which approaches to shared problems actually work?"

A system with genuine variation — where states can try different approaches to a problem and observe results — generates information that a uniform national policy cannot. California's air quality standards have driven automotive technology improvements that benefited the entire country, not because California was necessarily right in advance, but because California was able to try. Massachusetts's health insurance experiment informed the Affordable Care Act — again, not because Massachusetts had the final answer, but because Massachusetts could run the experiment.

This is not an argument against federal action. It is an argument about the epistemic value of variation. Some problems are genuinely national — monetary policy, national defense, interstate externalities, constitutional rights. Variation in how these problems are handled produces fragmentation without learning. But many public problems are local enough, and uncertain enough, that variation serves the goal of discovery.

What this means for practitioners

Local and state officials rarely think of their policy choices in these terms. They are not running experiments for the benefit of the national evidence base. They are trying to solve problems with limited resources and political constraints.

But the framing matters for institutional design. A city that pilots a new approach to library fines, documents the results, and publishes the findings — whether positive or negative — is contributing to a commons. A state that tries a new enrollment process, measures completion rates before and after, and makes the results available to other states has generated evidence that would not otherwise exist.

The constitutional structure creates the space. The institutional culture determines whether that space is used for experimentation and learning, or whether it produces variation without knowledge.

Brandeis's insight was not that states should be left alone. It was that the federal structure, at its best, is a system for discovering what works — one that uses the diversity of the country as a resource rather than a problem to be solved.

The laboratories metaphor has been domesticated into a political slogan. The underlying idea is more demanding and more interesting: that good governance requires not just making decisions, but learning from them.

Next issue: Default effects and the persistence of behavioral change — when does opt-out enrollment produce durable commitment, and when does it produce grudging compliance?