Chapter 4 — Federalism: The Endless Negotiation Between National and State Power

"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." — Justice Louis Brandeis, New State Ice Co. v. Liebmann (1932), dissenting

"States' rights" — a phrase used at different moments by Thomas Jefferson, John C. Calhoun, Strom Thurmond, the California legislature passing sanctuary laws in 2017, the Texas legislature passing abortion restrictions in 2021, and progressive city councils declaring themselves Second Amendment sanctuaries in 2023.

1. Two highways, one country

In October 2024, a 28-year-old Texas resident — call her Maya — discovered she was pregnant. She did not want to be. Texas had banned abortion after roughly six weeks since 2022 and had banned it almost entirely after the Supreme Court's Dobbs v. Jackson Women's Health Organization decision returned the question of abortion regulation to the states. Maya looked up her options. The closest legal abortion provider was in Albuquerque, New Mexico — about a 12-hour drive. She borrowed a car from a friend, took two days off her hourly job at a Houston warehouse, drove west on I-40, crossed the state line into New Mexico, received care at an out-of-state clinic, drove home, and went back to work on Monday. She was not breaking any law. New Mexico, the state where she received care, had no significant restrictions; Texas, the state where she lived, had nearly total prohibition. The state line — an arbitrary geographic boundary running roughly along longitude 103° west — determined which medical procedure she could legally obtain.

In the same month, a 22-year-old Colorado college student — call him Tyler — drove east on I-70 with two pre-rolled marijuana joints in his glove compartment that he had purchased legally at a licensed Denver dispensary. About 250 miles into his drive, he crossed into Kansas. The instant his car entered Kansas, those same two joints became contraband under both state and federal law. Tyler did not know he had committed a crime. He had committed two: a Kansas misdemeanor and a federal Schedule I drug-possession violation. He arrived at his grandmother's house in Wichita without incident. The federal government, as a matter of declared policy since the Cole Memo of 2013 (rescinded in 2018, partially restored under various administrations since), generally does not enforce federal marijuana law against personal use in states that have legalized it — but the law remains on the books, the products move across state lines as a matter of constant practice, and the legal status of any given joint depends on which side of an invisible line it sits on.

These are two ordinary American stories from the 2024–2025 period, and they encapsulate the core puzzle of this chapter. The United States is a single country with one currency, one military, one language of government, and a single Supreme Court whose opinions bind every state. It is also a country in which, depending on which state line you cross, the same conduct can be a felony or a constitutional right, the same medical procedure can be a federally protected service or a state-level crime, and the same employer can owe you a $7.25 minimum wage or a $16.50 minimum wage. The American system was built on the proposition that this is a feature, not a bug — that the right answer to many political questions is "let the states decide" — and the system has been arguing about how literally to take that proposition for 235 years.

This chapter is about that argument. It is not a static description of "powers reserved to the states." It is an account of how the boundary between national and state authority has been redrawn, repeatedly, by constitutional text, by Supreme Court interpretation, by partisan political alignment, by money, by war, and by the simple practical pressure of a growing country. We are now living in what we will call the asymmetric or policy-package era of federalism — an era in which red states and blue states have diverged on policy so substantially that the country increasingly behaves like two distinct legal regimes loosely federated by a shared currency, a shared military, and a shared Supreme Court. Whether this is alarming or healthy depends on your priors about national uniformity. We will treat both possibilities seriously.

By the end of this chapter you should be able to:

1. Identify the constitutional sources of national and state authority, and explain which clauses do which work.
2. Recount the four broad eras of American federalism — dual, cooperative, new, and asymmetric — and identify the cases and political moments that mark the transitions.
3. Explain the major mechanisms of federal-state interaction: enumerated powers, implied powers, preemption, the spending power, and anti-commandeering.
4. Read a federal grant arrangement (matching, block, categorical, conditional) and understand who actually controls the money.
5. Apply federalism reasoning to current controversies including marijuana policy, sanctuary jurisdictions, election administration, post-pandemic public health, and post-Dobbs abortion regulation — recognizing that federalism arguments cut whichever way they cut, and that both partisan camps have learned to use them.

2. The constitutional foundations

The U.S. Constitution does not contain the word "federalism." It contains, instead, a structure that has to be inferred from a handful of clauses spread across Articles I, IV, and VI, and the Tenth and Fourteenth Amendments. The four most important load-bearing clauses are:

Article I, Section 8 — the enumerated powers of Congress. Article I §8 is the constitutional list of things the federal government is allowed to do. It is a list — not a general grant. The list includes: laying and collecting taxes, regulating commerce among the several states (the Commerce Clause), coining money, declaring war, raising and supporting armies, establishing post offices, and a final catch-all permitting Congress "to make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers" (the Necessary and Proper Clause, sometimes called the Elastic Clause).

In plain English: Congress can do the specific things in this list, and anything else the list logically implies. The Founders thought this was a closed set. The Supreme Court, since 1819, has consistently interpreted "necessary and proper" to be quite generous — and the Commerce Clause, in particular, has done astonishing amounts of constitutional work. The federal minimum wage, civil-rights public-accommodations law, federal drug enforcement, federal environmental protection, the Affordable Care Act's individual mandate (in part), and most other federal regulation of private economic life ultimately rests on Congress's authority to regulate "commerce among the several states."

Article VI, Clause 2 — the Supremacy Clause. When state and federal law conflict, federal law wins. The Constitution and federal statutes "made in pursuance thereof" are "the supreme Law of the Land," and judges in every state are bound by them, "any Thing in the Constitution or Laws of any State to the Contrary notwithstanding." This is the clause that ended secession debates as a constitutional matter (whatever the Civil War settled with rifles, the Supremacy Clause settled with text). It is also the clause that powers the doctrine of preemption, which we will meet in §4.

The Tenth Amendment. "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." This is the constitutional anchor for the proposition that the federal government is one of enumerated powers — it can do what the Constitution lets it do, and the rest is for states (or for individuals, the "or to the people" tail). For most of the 19th century this clause was read with real teeth. From roughly 1937 to 1995 it was read as a near-truism with little independent force ("if the federal government has the power, it has the power; if it doesn't, it doesn't, and the Tenth Amendment doesn't add anything"). Since 1995 it has, again, been read with somewhat more bite — though how much more is one of the most-litigated questions in modern constitutional law.

The Fourteenth Amendment, Section 5. Ratified in 1868, the Fourteenth Amendment is the most important constitutional change to the federal-state balance in American history. Section 1 forbids states from denying "the equal protection of the laws" or due process to "any person." Section 5 grants Congress the power to enforce these provisions through "appropriate legislation." Most modern federal civil-rights law — the 1964 Civil Rights Act, the 1965 Voting Rights Act, the Americans with Disabilities Act, the Fair Housing Act — rests on Section 5 (sometimes alongside the Commerce Clause). The Fourteenth Amendment is the textual basis for the proposition that there are limits to what states may do to their own residents that the federal government will enforce.

These four pieces of text, plus the Treaty Power (Article II), the General Welfare Clause (often paired with the Spending Clause in Article I §8), and a handful of others, form the constitutional skeleton of American federalism. The flesh on that skeleton is two centuries of Supreme Court interpretation, congressional legislation, presidential practice, and political negotiation. We turn to that next.

3. Four eras of American federalism

It is useful to think of American federalism as having gone through four distinct phases. The transitions between them are messier than this schema implies, and historians who specialize in any one period will object to the simplifications. But for orientation:

3.1 Dual federalism, 1789–1937

For roughly the first 150 years of the Republic, the Supreme Court generally treated the federal and state governments as occupying separate, non-overlapping spheres — what scholars later named "dual federalism" or "layer cake" federalism. The federal government did the things on the Article I §8 list. The states did everything else: contract law, family law, criminal law, education, most economic regulation. The boundary between them was conceived as relatively clear and policed by the courts.

The defining cases of this era pull in different directions, which is part of what makes the period interesting:

McCulloch v. Maryland (1819). Maryland imposed a tax on the Second Bank of the United States, a federally chartered bank, hoping to drive it out of the state. James McCulloch, the bank's Baltimore branch cashier, refused to pay. Chief Justice John Marshall, writing for a unanimous Court, decided two issues. First, did Congress have the constitutional power to charter a bank, when the Constitution did not specifically authorize one? Yes — under the Necessary and Proper Clause, Congress could choose any means rationally related to a legitimate enumerated power. ("Let the end be legitimate, let it be within the scope of the Constitution, and all means which are appropriate, which are plainly adapted to that end, which are not prohibited, but consist with the letter and spirit of the Constitution, are constitutional.") Second, could Maryland tax it? No — "the power to tax involves the power to destroy," and a state cannot destroy a federal instrumentality. McCulloch is the case that makes the federal government larger than its enumerated list. Every subsequent expansion of federal power runs through McCulloch's reasoning.

Dred Scott v. Sandford (1857). The other side of dual federalism. Chief Justice Roger Taney, writing for a 7–2 Court, held that Black Americans of African descent — whether enslaved or free — could not be citizens of the United States, and that Congress lacked the power to ban slavery in the territories. The decision was a constitutional disaster (it helped precipitate the Civil War) and was effectively repudiated by the Thirteenth and Fourteenth Amendments. But it reflected the dual-federalism logic: Congress's powers were enumerated and limited, the rest belonged to states (or to "the people"), and the Court read both with a fierce strictness — when it suited the Court's politics.

Lochner v. New York (1905). At the height of the Industrial Revolution, the Court read the Fourteenth Amendment's due-process clause to protect a substantive "liberty of contract," striking down a New York law that limited bakery work to 60 hours per week. Lochner itself was about a state regulation, but it stood for a broader proposition: economic regulation, whether state or federal, was constitutionally suspect. The Court would strike down minimum-wage laws, child-labor laws, and assorted New Deal measures in the same vein for the next thirty years. Lochner is now almost universally regarded as wrongly decided — among scholars across the political spectrum — though the question of why it was wrong (overreach by the Court? overreach by the substantive-due-process doctrine? wrongly identified rights?) is itself contested, and figures into modern debates about Roe (1973), Casey (1992), Dobbs (2022), Obergefell (2015), and the proper scope of unenumerated rights.

What ended the dual-federalism era was a combination of the Great Depression, the New Deal, and a constitutional crisis in 1937 in which Franklin Roosevelt threatened to "pack" the Supreme Court. In the spring of that year, two justices switched their votes (the famous "switch in time that saved nine"), and the Court began upholding federal economic regulation under the Commerce Clause that it had been striking down. Whether this was good constitutional law or bad — whether the Court capitulated to political pressure or finally read the Constitution correctly — is a debate among constitutional scholars to this day.

3.2 Cooperative federalism, 1937–1980

From 1937 onward, the Supreme Court read the Commerce Clause and the Spending Clause expansively, and the federal government grew steadily into domains that had previously belonged to the states. The metaphor changed from layer cake (separate spheres) to marble cake (intertwined spheres) — federal and state governments increasingly cooperated on programs that neither could fully run alone. Highways, education, public health, environmental protection, welfare, and after 1965 health care all became joint federal-state enterprises, with federal money flowing to states subject to federal rules.

The defining case is Wickard v. Filburn (1942). Roscoe Filburn was an Ohio farmer who grew about 12 acres of wheat more than his federal allotment. He used the excess wheat himself — to feed his own livestock and bake his own family's bread. He never sold a single grain of it across state lines. Could the federal government, under the Commerce Clause, fine him for growing too much wheat? The Court said yes. By growing his own wheat, Filburn reduced the amount he would otherwise have bought on the market; the cumulative effect of many farmers like Filburn doing the same thing affected interstate commerce; and Congress could regulate any activity, however local, whose aggregate effect substantially affected interstate commerce. Wickard is the high-water mark of Commerce Clause expansionism. After Wickard, the question "is this within the federal power?" almost always answered itself — yes.

This era also produced civil rights legislation that fundamentally reset the federal-state relationship. The 1964 Civil Rights Act used the Commerce Clause to ban racial discrimination in public accommodations (hotels, restaurants, theaters that served interstate travelers). The 1965 Voting Rights Act used the Fifteenth Amendment to require federal preclearance of voting changes in covered jurisdictions. The Court upheld both. Whatever else cooperative federalism meant, it meant that the federal government could now tell states they could not maintain Jim Crow.

3.3 New federalism, 1980–2008

Beginning with the Reagan administration and accelerating in the 1990s, two things changed. First, Republican administrations began consciously pushing power back to the states through block grants, regulatory rollbacks, and welfare reform. Second, the Supreme Court, under Chief Justice William Rehnquist, began policing federal power again — for the first time in nearly sixty years. This phase is sometimes called the "Federalism Revolution" or the era of New Federalism.

Garcia v. San Antonio Metropolitan Transit Authority (1985) is often the marker chosen for the end of one trajectory, not its beginning. The case held, 5–4, that Congress could apply federal minimum-wage and overtime rules to state employees — and that the political process, not the courts, should police federalism. Justice Harry Blackmun, writing for the majority, said in essence: the states have senators, they have representatives, they can defend their own interests through politics. The Tenth Amendment was again read as nearly a truism. Garcia looked like the final triumph of cooperative federalism.

It was not. A decade later, in United States v. Lopez (1995), the Court for the first time since 1937 struck down a federal statute on Commerce Clause grounds. The Gun-Free School Zones Act made it a federal crime to possess a firearm within 1,000 feet of a school. Chief Justice Rehnquist, writing for a 5–4 majority, held that gun possession near a school was not "economic activity" and had no plausible substantial effect on interstate commerce, and Congress could not regulate it under the Commerce Clause. Lopez surprised observers — it was the first such ruling in three generations — and signaled that the Court was willing to enforce some outer limit on federal power.

Printz v. United States (1997) followed. Congress had required state and local law-enforcement officers to perform background checks on prospective handgun purchasers under the Brady Bill. The Court, again 5–4, held that the federal government cannot "commandeer" state officials to administer a federal regulatory program. This is the anti-commandeering doctrine: Congress can offer states money to run programs, can preempt state law, and can regulate states alongside private actors when they engage in conduct Congress has the power to regulate, but it cannot simply order state officials to enforce federal law.

By the Bush II era (2001–2009), the basic shape of New Federalism was: a federal government that did much more than the original Constitution contemplated, but that the Court would occasionally cabin in cases where the federal claim was particularly weak.

A fourth case worth flagging from this era is Gonzales v. Raich (2005), which complicates the New Federalism story considerably. Angel Raich, a California resident with serious medical conditions, used marijuana that her caregivers grew in California, in compliance with California's 1996 medical marijuana law. No money changed hands, and nothing crossed a state line. Could the federal government prosecute her under the Controlled Substances Act? The Court — with Justice Antonin Scalia in the majority and Justice Sandra Day O'Connor in a vigorous dissent that quoted Brandeis on laboratories — said yes. Personal medical use of homegrown marijuana was, like Filburn's homegrown wheat, the kind of activity that, in the aggregate, substantially affected interstate commerce, and Congress could regulate it. Raich told observers that Lopez and Morrison had not actually narrowed the Commerce Clause much; they had carved out a small zone for non-economic activity (gun possession, gender-motivated violence as a tort) but had left the broad post-Wickard logic intact for activity that could be characterized as economic. New Federalism was thus more rhetorical than doctrinal in many of the years it was supposed to have governed.

3.4 Asymmetric federalism, 2008–present

The current era does not yet have a settled name. We will call it asymmetric federalism or policy-package federalism, following scholars like Jessica Bulman-Pozen and Heather Gerken. Its defining features:

1. Partisan sorting at the state level. Through the 1990s, "trifecta" states — those where one party controlled the governorship and both legislative chambers — were a minority. By 2025, roughly 38 of 50 states had unified party control of state government (NCSL data). State politics has become more like national politics — sorted, partisan, ideological — and the policy distance between Texas and California, or Florida and Massachusetts, has grown rather than narrowed.

2. Policy divergence on culturally salient questions. Abortion access, gun regulation, marijuana legality, immigration enforcement, public-health authority, education curricula, transgender medical care for minors, climate policy, and minimum wage all now vary sharply by state.

3. Both parties using federalism arguments instrumentally. Conservatives, who through most of the post-Civil-War period favored states' rights as a check on federal power, have shifted to favor federal preemption when state policy moves left (federal preemption of state climate regulation; federal limits on state sanctuary policies; federal preemption of state vaccine rules). Progressives, who through most of the 20th century favored federal civil rights enforcement against recalcitrant states, have shifted to favor state and local autonomy when federal policy moves right (sanctuary cities, state climate compacts, state-level abortion shield laws). The intellectual gymnastics on each side are visible. Both teams have absorbed the lesson that federalism is a vehicle, not a destination.

4. Supreme Court reinforcement. The Roberts Court has reinforced asymmetric federalism, both deliberately and incidentally. NFIB v. Sebelius (2012) made Medicaid expansion optional, instantly creating a 50-state patchwork. Murphy v. NCAA (2018) struck down a federal sports-betting ban on anti-commandeering grounds, allowing states to legalize sports betting (now legal in 38+ states). Dobbs v. Jackson Women's Health Organization (2022) returned abortion to the states, producing the divergence Maya navigated in §1.

The asymmetric-federalism question is whether this is sustainable. We will return to it in §11.

4. Mechanisms

How does federalism actually work, mechanically? Five mechanisms do most of the day-to-day work.

4.1 Enumerated and implied powers

Congress's powers are listed in Article I §8. The catalog includes the power to lay taxes, borrow money, regulate commerce among the states and with foreign nations and Indian tribes, establish a uniform rule of naturalization, coin money, establish post offices, declare war, raise armies, and a handful of others. Anything genuinely outside this list, and outside the powers Congress has been granted by later amendments, is reserved to the states by the Tenth Amendment.

Implied powers are powers Congress has not because they are listed, but because they are "necessary and proper" to carry out the listed powers. McCulloch v. Maryland established the wide reading: any means rationally related to a legitimate end is permissible. The federal criminal code, the Federal Reserve, federal regulatory agencies, the Department of Education, NASA — none of these is in Article I §8. All of them rest on implied powers traced back to enumerated grants.

4.2 The Commerce Clause

Of all the enumerated powers, the Commerce Clause does the most work. After 1937 the Court read it broadly enough to encompass virtually all economic activity (Wickard, 1942). Heart of Atlanta Motel v. United States (1964) used it to require integration in public accommodations. Gonzales v. Raich (2005) used it to allow federal enforcement of marijuana prohibition against personal medical use legal under California law.

But the Commerce Clause is no longer infinite. Lopez (1995, gun possession near schools) and Morrison (2000, civil remedy for gender-motivated violence) struck down statutes as exceeding the commerce power. NFIB v. Sebelius (2012) held that Congress could not require Americans to purchase health insurance under the Commerce Clause — though it could impose a tax on those who did not, sustaining the individual mandate under the taxing power. The current rule of thumb: Congress can regulate economic activity (broadly defined) that has a substantial effect on interstate commerce in the aggregate. It cannot regulate non-economic activity simply by aggregating it. The line between economic and non-economic activity is not always clear, and lawyers fight over it constantly.

4.3 Preemption

When Congress validly legislates in a field, federal law can preempt state law. Preemption comes in three forms: express (the federal statute states it preempts state law — common in airline, drug, and energy regulation), conflict (compliance with both federal and state law is impossible, or state law obstructs the federal purpose), and field (Congress has so thoroughly occupied a field that no state regulation is permissible — immigration is the classic example).

Preemption is a major battlefield in modern federalism. Federal preemption of state tort lawsuits against drug manufacturers (Vaccine Injury Compensation, FDA-approved labeling) is contested. Federal preemption of state climate rules is contested. Federal preemption of municipal labor laws is contested. The instrumentality of preemption shifts over time: when states are doing things national majorities don't like, Washington reaches for preemption; when states are doing things national majorities like, Washington holds back.

4.4 The Spending Power and conditional grants

Congress may "lay and collect Taxes... to provide for the common Defence and general Welfare of the United States" (Article I §8, Clause 1). It may attach conditions to federal grants — and through those conditions effectively regulate areas it could not directly regulate. The seminal case is South Dakota v. Dole (1987), in which Congress conditioned a small percentage of federal highway funds on states adopting a 21-year-old minimum drinking age. The Court (Chief Justice Rehnquist) upheld the condition: Congress could not directly set the drinking age (a state matter) but could induce states to do so through the Spending Clause, so long as the condition was related to the federal interest, was clearly stated, and was not so coercive as to amount to "compulsion."

That last requirement got teeth a quarter century later. In NFIB v. Sebelius (2012), the Affordable Care Act conditioned all federal Medicaid funding (a substantial fraction of state budgets) on states' agreeing to expand Medicaid eligibility. The Court held this crossed the line from inducement to coercion: states could not realistically refuse, so the condition was unconstitutional. The remedy: the expansion became optional. As of 2025, 41 states (and D.C.) had expanded Medicaid; 9 had not. The expansion gap has been one of the largest sources of policy variation in American health care.

4.5 Anti-commandeering

The federal government cannot order states to enact or enforce a federal regulatory program. Printz v. United States (1997) struck down the requirement that local police perform Brady-Bill background checks. Murphy v. NCAA (2018) struck down a federal statute that prohibited states from authorizing sports gambling — Congress could ban sports gambling itself, but it could not forbid states from changing their own laws. The anti-commandeering doctrine has been one of the most consequential doctrines for modern federalism. It is the constitutional basis on which sanctuary jurisdictions can refuse to assist federal immigration enforcement, on which states could legalize marijuana while it remained federally illegal, and on which states retain meaningful policy discretion even when they accept federal money.

5. Fiscal federalism: where the money actually flows

You cannot understand modern American federalism without understanding the money. The federal government, through grants to states and localities, provides on the order of $1.1 trillion per year in transfers. For most state governments, federal money is not a small supplement — it is between a quarter and half of the total state budget.

A few representative figures (rounded; based on Treasury, OMB, and NASBO data 2023–2024):

The federal share of total state and local spending hovers around 30%. In a state like Mississippi or Louisiana, federal money can be 45% or more of the state budget; in a state like Massachusetts or California (with large state tax bases), federal money is closer to 25%. Federal money is the most important policy lever the national government has, because the conditions attached to it can drive states to adopt federal priorities even where Washington has no direct regulatory power.

There are four common types of federal grants:

1. Categorical grants — money for a specific purpose, with detailed federal rules (e.g., Title I funding for schools serving low-income students). Most federal grant money is categorical.
2. Block grants — money for a broad purpose, with substantial state discretion (e.g., the Community Development Block Grant, the TANF block grant). The Reagan and Trump administrations both proposed converting categorical programs to block grants; this generally reduces federal control and reduces total funding (block grants tend not to be inflation-adjusted or expanded as program need grows).
3. Matching grants — federal money conditioned on a state contribution. Medicaid is the largest. The federal "match" varies by state wealth — Mississippi's federal share (the Federal Medical Assistance Percentage, FMAP) is around 77%, while New York's is the statutory floor of 50%. Matching grants are politically powerful: every state dollar leverages federal dollars, so states that don't participate are leaving money on the table.
4. Mandates (funded and unfunded) — federal requirements that states must comply with. The Americans with Disabilities Act imposes substantial costs on state and local governments without commensurate federal funding (though there is some). The Unfunded Mandates Reform Act of 1995 required Congress to estimate mandate costs but did not, in practice, prevent unfunded mandates.

A practical example: when Washington wanted to lower the drinking age to 21 in the 1980s, it could not directly regulate alcohol sales (that's a state power, since the Twenty-First Amendment). What it could do — and did — was condition 5% of federal highway funds on states' adopting a 21 drinking age. Within a few years every state had complied. South Dakota v. Dole (1987) upheld the arrangement. The Spending Power, in practice, often does what direct federal regulation cannot.

6. States as laboratories: the successes and the failures

Justice Brandeis's "laboratories of democracy" formulation, from a 1932 dissent quoted at the top of this chapter, has become the standard defense of federalism: states can experiment, and the rest of the country can learn. The framing is partly true and partly aspirational. Some real examples of state-level experimentation that the country has learned from — and some examples of experiments that produced harm, that other states refused to copy, or that revealed the limits of state-level innovation.

Successes (or at least durable diffusions):

• Massachusetts and the individual mandate (2006). Republican Governor Mitt Romney signed "Romneycare," a state insurance reform requiring all residents to obtain health insurance (with subsidies, an exchange, and an individual mandate). The Massachusetts experiment ran for four years before the Affordable Care Act borrowed its structure for the federal level — a Republican governor's policy becoming a Democratic president's signature legislation. Federalism worked: a state tried it, the country watched, and the policy was ported with adjustments.

• California and emissions standards. Under §209 of the Clean Air Act, California has, since 1970, been allowed to seek a federal waiver to set vehicle emissions standards stricter than federal. Other states may then adopt California's standards. This has driven national auto-industry behavior for decades — manufacturers don't generally make two cars (a "California car" and a "rest of country" car) when they can make one car compliant with both standards. We treat this in detail in Case Study 1.

• Wisconsin's "Wisconsin Works" (W-2) and the 1996 federal welfare reform. Republican Governor Tommy Thompson's 1996 state welfare reform, focused on work requirements and time limits, was the explicit model for the federal Personal Responsibility and Work Opportunity Reconciliation Act of 1996. Whether the federal reform was a success or a failure remains contested — caseloads dropped sharply; child poverty patterns are complicated — but the policy diffusion from state to federal level is unambiguous.

• Oregon's Death With Dignity Act (1997). Oregon was the first state to legalize physician-assisted dying for terminally ill adults. Other states (Washington, Vermont, California, Colorado, Hawaii, New Jersey, Maine, New Mexico, Montana via court) have since adopted similar laws. Federal law neither permits nor preempts; the matter is left to states.

• Marijuana liberalization. Beginning with California's medical-marijuana initiative in 1996 and Colorado/Washington's recreational legalization in 2012, state-level liberalization has progressed despite continued federal prohibition. By 2025, 24 states (plus D.C.) have legalized recreational marijuana; 38 have legalized medical marijuana. We treat this in §7.

Failures or worse-than-failures:

• Race to the bottom in some regulatory areas. States competing for businesses sometimes weaken regulations that produce externalities crossing state lines. Environmental regulation, labor regulation, and corporate governance have all seen variants of this dynamic. Delaware's corporate law dominance — a function of competing state legislatures eager to attract incorporation — is a classic example, as is the use of low-tax states as offshore-equivalent corporate domiciles.

• Nullification rhetoric. Through American history, states have intermittently asserted the power to "nullify" federal law they consider unconstitutional. The Kentucky and Virginia Resolutions (1798) advanced an early version. South Carolina's nullification of federal tariffs (1832) was beaten back by Andrew Jackson. Various Southern states attempted "massive resistance" to Brown v. Board of Education (1954). In recent years, "Second Amendment sanctuary" counties have declared they will not enforce federal gun laws, and "sanctuary" jurisdictions have declared they will not assist federal immigration enforcement. The constitutional reality, since Cooper v. Aaron (1958), is that states cannot lawfully nullify federal law. The political reality is that states can sometimes effectively decline to assist (anti-commandeering allows refusal to actively enforce), and they can sometimes use political pressure to influence federal enforcement priorities.

• State-level civil-rights regression. From the end of Reconstruction to the 1960s, state and local governments throughout the American South systematically denied Black Americans the protections of the Constitution. The "laboratories" frame can flatter into ignoring that some state-level experiments — Jim Crow, eugenic sterilization, anti-miscegenation laws — were experiments only in the sense that they were tried.

The honest reading of "laboratories of democracy" is: the federal-state structure does enable experimentation, the experimentation does generate diffusion and learning, and states can also do things that are bad. Whether on net the gains exceed the costs depends on the issue and the time period.

7. Marijuana federalism: a vivid current case

There is no clearer current illustration of how federalism actually operates than American marijuana policy.

The federal law. The Controlled Substances Act of 1970 places marijuana on Schedule I — drugs deemed to have no accepted medical use and high potential for abuse. Manufacturing, distribution, possession, and use of marijuana are federal crimes. As of late 2025, the Drug Enforcement Administration is in the late stages of a process initiated in 2024 to reclassify marijuana to Schedule III, but the process is incomplete and federal prohibition remains the legal baseline.

The state law. As of January 2025: - 24 states (plus D.C.) have legalized recreational marijuana for adults. - 38 states have legalized medical marijuana. - Roughly half the U.S. population lives in a state where adult recreational use is legal under state law. - The legal cannabis market has revenues in the range of $30–35 billion annually (BDSA, MJBizDaily estimates 2024).

The federalism puzzle. Federal law prohibits what state law allows. Under Article VI, federal law is supreme. So why isn't every state-licensed dispensary being shut down by federal agents?

Three reasons:

1. Anti-commandeering. The federal government cannot order California police to enforce federal marijuana law. California chooses not to enforce the federal prohibition; it enforces only its own (more permissive) state law. Federal agents can enforce federal law themselves — they have the personnel and authority — but the practical reality is that the DEA cannot replace state and local police for routine drug enforcement at scale.

2. Federal forbearance. The Justice Department's "Cole Memo" (2013, under Obama) instructed federal prosecutors not to prioritize enforcement against state-licensed marijuana businesses absent specific federal interests (sales to minors, sales across state lines, etc.). The memo was rescinded by Attorney General Sessions in 2018 but the actual prosecutorial pattern has remained largely consistent under both subsequent administrations. The Rohrabacher-Farr/Joyce amendment, attached to federal appropriations bills since 2014, prohibits the DOJ from spending money to interfere with state medical-marijuana programs. (Recreational programs are not similarly protected, but in practice federal enforcement against state-licensed recreational businesses has been minimal.)

3. Banking complications and the practical mess. Because marijuana remains federally illegal, federally regulated banks cannot serve marijuana businesses without potential federal liability. The cannabis industry is therefore largely unbanked or quasi-banked — operating in cash, using state-chartered credit unions, or relying on workarounds. The SAFE Banking Act, designed to fix this, has passed the House multiple times since 2019 but has not become law as of 2025. Sales tax revenues from legal cannabis (over $4 billion in 2023 across legal-recreational states) flow to state coffers, while federal taxes apply normally — under §280E of the Internal Revenue Code, marijuana businesses cannot deduct ordinary business expenses (because the business is federally illegal). The result is an industry paying effective federal tax rates around 70%.

The marijuana situation is constitutionally untidy and practically functional. Tyler from §1 — who carried two joints across a state line — committed a federal crime that, in nearly all real-world circumstances, no one will ever prosecute. Gonzales v. Raich (2005) confirmed that the federal government has the constitutional power to enforce its prohibition against personal medical use even in states where it is legal. Since 2005, with rare exceptions, the federal government has chosen not to. The Supremacy Clause is real; so is prosecutorial discretion; so is the political cost of overriding 24 state legislatures and a majority of the public.

This is federalism at its most practically improvisational — a long-running negotiation between federal authority that exists on paper and state policy that exists in practice.

8. Sanctuary jurisdictions and immigration federalism

Immigration is a federal power. The Constitution gives Congress authority over naturalization (Article I §8 Clause 4), and the Supreme Court has long held that the federal government has plenary authority over immigration generally (Chae Chan Ping, 1889; Arizona v. United States, 2012). States cannot make their own immigration laws or set their own deportation policies.

But — and here is where anti-commandeering does decisive work — the federal government cannot order state and local police to enforce federal immigration law. The major federal mechanisms of cooperation are:

• 287(g) agreements. A provision of the 1996 immigration law allows local jurisdictions to opt into agreements giving local officers federal immigration-enforcement authority. Roughly 130 jurisdictions have such agreements as of 2025, mostly in conservative states.
• Detainer requests. Immigration and Customs Enforcement (ICE) can ask local jails to hold a person in custody beyond their state-law release date, so ICE can take custody. Since Galarza v. Szalczyk (3d Cir. 2014) and similar rulings, jurisdictions have been on notice that compliance with detainers is voluntary, not mandatory, and that improperly holding someone past their release date can create state-law liability for the locality.
• Information sharing. ICE and local jails generally share booking data through federal systems (Secure Communities, etc.). Some sanctuary jurisdictions limit this sharing.

A "sanctuary" jurisdiction is, in practice, a city, county, or state whose policies limit cooperation with federal immigration enforcement — refusing detainer requests, declining to share information beyond what federal law requires, declining to permit local resources to be used for civil immigration enforcement, and so on. The exact policies vary widely. Sanctuary policies emerged in the 1980s in religious congregations, expanded to cities, and have proliferated since 2017. As of 2025, 11 states and over 200 localities have some form of sanctuary policy.

The federalism picture has notable inversions. In the 1950s and 1960s, the federal government compelled Southern states to accept civil-rights enforcement, and those states invoked "states' rights" to resist. Today, the federal government attempts to compel progressive cities to accept federal immigration enforcement, and those cities invoke state and local autonomy to resist. The constitutional doctrine — that the federal government cannot commandeer state officials — is the same. The political valence has flipped.

The federal government has tried to penalize sanctuary jurisdictions through conditional spending. The Trump administration in 2017 attempted to condition federal law-enforcement grants on cooperation with ICE; multiple federal courts struck this down (the conditions were not authorized by the underlying statute, and may have run afoul of NFIB's coercion limit). Subsequent administrations have used different mechanisms with varying success. The legal questions are largely unsettled and likely to remain a live area of litigation.

A clean steel-man on each side:

• Critics of sanctuary policies argue: Federal immigration law is law; selective non-cooperation by progressive jurisdictions undermines the rule of law generally and protects individuals who have committed serious crimes (selectively framed). The political-accountability argument that sustained civil-rights enforcement against Southern resistance in the 1960s should sustain immigration enforcement against urban resistance now.

• Defenders of sanctuary policies argue: Local police effectiveness depends on cooperation from immigrant communities, who will not report crime or testify if they fear deportation. Anti-commandeering doctrine is the same doctrine that has, throughout American history, allowed states to be policy laboratories on contested questions. And the federal government's expansive use of immigration enforcement for what amount to civil violations — not criminal ones — outruns what local taxpayers have agreed to support.

Both arguments are real. We will return to immigration more fully in Chapter 31.

9. Election administration federalism

Article I §4 of the Constitution gives states the power to set the "Times, Places and Manner" of holding congressional elections, "but the Congress may at any time by Law make or alter such Regulations, except as to the Places of chusing Senators." This compromise — states run elections, Congress can override — has produced one of the most decentralized election systems in the democratic world. The United States has approximately 10,000 separate election jurisdictions (counties, cities, townships) running federal elections. Voter registration, ballot design, polling places, voting machines, mail-ballot rules, ID requirements, drop boxes, recount procedures, and certification rules vary by state and often by county.

A few representative variations as of 2024:

• Mail voting. Eight states conduct elections almost entirely by mail; thirty-some states allow no-excuse absentee voting; a handful require an excuse to vote absentee.
• Voter ID. Some states require a government-issued photo ID at the polls; others accept a signature; others have no ID requirement.
• Felony disenfranchisement. Maine and Vermont allow incarcerated people to vote. Most states restore voting rights upon release (or upon completion of parole). A handful require a separate civil-rights restoration.
• Same-day registration. About 23 states allow voters to register on Election Day; the rest have registration deadlines ranging from 7 to 30 days before.
• Drop boxes. Some states have hundreds; some require one per county; some have effectively none.

Federal law sets a floor through the Voting Rights Act of 1965 (most provisions still in effect; the preclearance provision struck down in Shelby County v. Holder, 2013), the National Voter Registration Act ("Motor Voter," 1993), and the Help America Vote Act (2002). Congress has the constitutional authority to set far more uniform rules than it has used, and proposals to do so (the For the People Act, the Freedom to Vote Act) have repeatedly failed in the Senate. Election administration remains overwhelmingly a state matter, with all the variation that produces. Chapters 22 and 36 develop the full picture.

10. Pandemic federalism: COVID-19 as stress test

The COVID-19 pandemic of 2020–2022 was a stress test of American federalism, and the results were mixed. A few observations:

The federal government had limited authority over many of the most consequential decisions. Public-health authority — quarantine, business closures, school closures, mask requirements — is mostly state and local. The federal government runs the FDA (vaccine and treatment authorization), CDC (surveillance and guidance), and FEMA (emergency response logistics), and it controlled the bulk of the financial response, but the policy decisions about whether to close schools, whether restaurants could open at 50% capacity, whether you needed a mask in a grocery store — those were almost all state and local. Florida and California adopted very different pandemic-era policies and produced different outcomes; comparing them is methodologically tricky (different demographics, different population densities, different industry mixes, different baseline rates of comorbidity), but the cross-state variation was real and occasionally striking.

Some specific federalism dynamics:

• Federal preemption questions. Could the federal government require employers above a certain size to mandate vaccination of employees? OSHA tried, under an emergency temporary standard. NFIB v. OSHA (2022) said no — the agency had exceeded its statutory authority. The Court did, however, uphold a parallel mandate for healthcare workers at facilities receiving Medicare/Medicaid funds, under the Spending Power.
• State countermeasures. Florida and Texas passed laws prohibiting employer vaccine mandates. Other states required them. Workers and employers found themselves operating under conflicting rules depending on jurisdiction.
• Vaccine distribution. The federal government procured and allocated vaccines; states and localities ran the actual distribution, with wide variation in efficiency.
• School policy. District-by-district decisions about reopening produced an extraordinary patchwork. Some students returned to in-person school in fall 2020; others did not return until fall 2021. The achievement effects of these variations are still being studied (the National Bureau of Economic Research and the Brown CEPR have produced extensive analyses).

A useful way to read pandemic federalism: it revealed both the strengths of decentralization (faster local adaptation, learning across jurisdictions) and the costs (inconsistency, public confusion, the inability to coordinate when coordination would have saved lives). Whether on net the system performed well or poorly depends on which counterfactual you are comparing it to. Reasonable observers across the political spectrum disagree.

11. The asymmetric-federalism era: where we are now

We have used the term asymmetric federalism to describe the current era. The defining feature is policy divergence among states, sustained by partisan sorting and reinforced by Supreme Court doctrine that has returned increasing numbers of contested questions to the states.

A few reference points for the divergence as of 2025:

• Abortion access. 13 states have near-total bans; 7 states have substantial restrictions; 30 states permit abortion under varying conditions. Median driving distance to a clinic from a banned-state center: ~250 miles. We treat this in Case Study 2.
• Gun regulation. Concealed-carry permitting ranges from "may issue" (rare in remaining states post-Bruen, 2022) to "shall issue" to constitutional carry (no permit required). Universal background checks are required in 21 states; not in 29. Magazine-capacity limits exist in 9 states.
• Marijuana. 24 recreational legal, 38 medical legal, 12 fully prohibited.
• Minimum wage. Federal floor: $7.25 (unchanged since 2009). State floors range from $7.25 to $17 (D.C., Washington, California). City floors go higher (Seattle, San Francisco, NYC ~$17–19).
• Public-school curriculum. Substantial state variation in how American history, racial history, and gender topics are taught. 18 states have laws restricting certain race-related instructional materials; several states have laws protecting it.
• Climate policy. California, New York, Washington, and a coalition of others coordinate ambitious state-level climate programs; Texas, Wyoming, West Virginia coordinate state-level resistance to federal climate regulation.
• Healthcare. Medicaid expansion: 41 states (and D.C.) yes, 9 no. State-level "public option" plans: a handful. Cross-state variation in uninsured rates ranges from ~3% (Massachusetts) to ~16% (Texas).

The argument that this divergence is a problem comes from across the political spectrum. From the left: fundamental rights should not depend on state of residence; the same constitutional country should not have a $7.25 minimum wage in some places and a $17 minimum wage in others; abortion is a basic medical right. From the right: border security and the federal-immigration-law structure are nationally important and being undermined by state defiance; vaccine mandates and pandemic policies needed national coordination; some state policies functionally export externalities. The shared anxiety: that the country is splintering into two policy regimes that have less and less in common, that this is bad for national unity, and that increasing national-level conflict is the natural consequence.

The argument that the divergence is healthy also comes from across the political spectrum. From the left: the federal level is unreliable, and state and local progressive governance produces real wins (climate, minimum wage, access); without state autonomy, the rightward turn of national politics in 2024–2026 would be unchecked. From the right: the federal level has overreached for decades; state autonomy on cultural and economic questions returns power to communities, and federalism is the constitutional design choice that has always allowed the country to live with deep disagreement; better disagreement-with-distance than enforced uniformity-with-conflict.

Both sides have learned to use the same tools. When a state pursues policy a national majority supports, federalism is "letting states experiment." When the same state pursues policy a national majority opposes, federalism is "states getting away with too much." The same is true of state-versus-federal preemption arguments, of states-as-laboratories framings, of "states' rights" rhetoric. The most honest description of contemporary American federalism is that it is morally and politically symmetric — it cuts whichever way it cuts. The Texas legislator who praises federalism for permitting Texas's six-week abortion ban and condemns sanctuary cities for refusing to assist ICE, and the California legislator who praises federalism for permitting California's emissions standards and condemns Texas's restrictions, are both being instrumentally consistent. The intellectual question is whether there is a principled commitment to federalism beyond its partisan utility — and there are scholars and politicians who have such a commitment, but they are rarer than the constitutional rhetoric suggests.

A question to sit with: Is asymmetric federalism a stable equilibrium, or a transitional phase? Three possibilities, each defended by serious scholars:

1. Stable equilibrium. Deep state-level policy variation is fine; the country has lived with substantial regional divergence throughout its history (Jim Crow North vs. South; rural West vs. urban East; Bible Belt vs. coasts), and the current divergence, however unprecedented in some specifics, is not categorically different. Federalism does its constitutional work by absorbing disagreement.

2. Transitional phase toward national resolution. Both pro-life and pro-choice advocates after Dobbs are pursuing national legislation (a federal abortion ban from one direction, federal codification of Roe from the other). Both gun-rights and gun-control advocates pursue national legislation. The current state-level patchwork is a battlefield, not a settlement; eventually national majorities will resolve these questions one way or another.

3. Transitional phase toward separation. The country is sorting so thoroughly that genuine policy union is becoming impractical. Some scholars have begun to write about a "soft secession" or "two Americas" outcome, where federal authority shrinks to a narrow military-and-currency core and most policy moves to the state level. This is the most pessimistic reading; it is also the one most resisted by those who believe the constitutional union is more durable.

Which of these three you find most plausible probably reflects your priors about American politics. The honest analytical answer is: we don't know yet. Federalism, in its asymmetric phase, is being tested in real time, and the resolution will play out over decades, not months.

A few more analytical observations worth holding onto as you read the rest of this book:

Federalism interacts with rights in unstable ways. When a federal court declares a federal constitutional right (the right to interstate travel, free speech, due process), states are bound. When a federal court declares that a question has no federal constitutional answer — which is part of what Dobbs did with abortion — states are free to differ. This means that the answer to "what does federalism let states do?" depends fundamentally on what the Supreme Court has said about which questions are constitutionally settled and which are not. A future Court that revisits, for example, Obergefell v. Hodges (2015, same-sex marriage) would rearrange the federalism map of marriage in a single decision. The Roberts Court's pattern in this regard — declining to nationalize new rights and occasionally returning previously nationalized ones to the states — is producing the asymmetric pattern. A different Court would produce a different one.

Federalism interacts with money in stabilizing ways. The federal government's fiscal leverage is often the decisive variable in whether state policy diverges or converges. States that take Medicaid expansion are operating under federal rules; states that reject it are not. States that take federal highway money operate under federal speed-limit and drinking-age rules. The Spending Power is the most consistently effective tool the federal government has for inducing state cooperation, and unlike preemption, it does not raise anti-commandeering concerns (states are nominally free to refuse). This is why proposals to dramatically expand or contract the federal grant system are, at heart, federalism proposals — change the money flow and you change the practical balance of national and state authority more than any constitutional amendment plausibly would.

Federalism interacts with partisan identity in ways the Founders could not have anticipated. The Founders designed a system in which states would be loci of competing interests; they did not foresee that two national parties would sort the states into red-and-blue blocs that would behave more like coalitional partners than independent political actors. This is the crucial novelty of the asymmetric era and the reason older textbook treatments of federalism — which describe states as varied independent actors pursuing their own interests — feel out of touch with current political reality. Most state policy decisions in the most ideologically charged areas now line up cleanly with national party programs. State-level resistance to federal authority is partisan resistance; state-level innovation is often partisan innovation. This is not the dual-federalism world Madison and Hamilton debated.

The honest summary: American federalism is a working constitutional structure, an inherited institutional pattern, an active political tool, and a chronic source of legal uncertainty. It is the connective tissue of American government — which is why it appears in this textbook as a foundations chapter, and why it will appear, sometimes by name and sometimes only structurally, in nearly every chapter that follows.

12. Where we go from here

This chapter has set up federalism as the constitutional structure within which most of the rest of American government operates. Federalism is not just one chapter of the textbook — it is the connective tissue that ties together Civil Rights (Ch 6), Congress's spending power (Ch 9, 11), the lower federal courts (Ch 13–14), state and local government (Ch 15), the policy chapters in Part IV (Ch 27–32), and the chapters on voting, gerrymandering, and democratic erosion in Part V (Ch 35–37). When you read those later chapters, ask: which level of government is acting here, and on what authority? What can the other level do to resist or reinforce? Where does the money come from? The federalism lens does an enormous amount of explanatory work.

Two specific previews:

• Civil rights (Ch 6). The story of American civil rights is a federalism story. Reconstruction was about federal authority over recalcitrant states; the post-Reconstruction retreat was about returning power to the states; the civil-rights movement was about federal reassertion via the Fourteenth Amendment, the Commerce Clause, and the Voting Rights Act; post-Shelby County (2013) is about a partial retreat. The federalism question and the civil-rights question are not separable.

• Voting rights and election administration (Ch 22, 36). The American election system's decentralization is a federalism choice with major substantive consequences. The variation in voting rules across states is one of the most consequential expressions of federalism in modern American politics, and the question of whether to nationalize election administration is a live federalism question.

The endless negotiation continues. The next chapter turns to civil liberties — the limits the Constitution places on what any government, federal or state, can do to individuals. Federalism tells us which government can act; civil liberties tell us what it cannot do regardless. Together, they form the structural and rights frames within which all American politics operates.

Sources and data referenced in this chapter include the Office of Management and Budget, the Treasury Department's State and Local Government Series, the Kaiser Family Foundation's State Health Facts, the NCSL legislative database, the Brennan Center's elections tracking, the Marijuana Policy Project's state-by-state tracker, the Guttmacher Institute's abortion-policy tracker, the National Association of State Budget Officers (NASBO), and the supreme court opinions cited by name. Specific statistical figures are accurate as of late 2025; readers should consult primary sources for current data.