Chapter 15: State and Local Government — Where Most Government Actually Happens

"The states are the laboratories of democracy." — Justice Louis Brandeis, New State Ice Co. v. Liebmann (1932), dissenting

"All politics is local." — Tip O'Neill, attributed (the line predates him; he made it famous)

Why this chapter is here

Open any standard American Government textbook and count the chapters on Congress, the presidency, the federal courts, and the federal bureaucracy. Then count the chapters on state and local government. The federal side wins by roughly five to one, sometimes ten to one.

Now do a different exercise. Walk through your week. The road you drove on this morning — who paved it, who fills the potholes, who decides where the stop signs go? The water that came out of your tap — who runs the pipes, who tests it for contaminants, who sets the rates? The school across the street — who hires the teachers, who picks the textbooks, who decides whether to close on snow days? The police car that drove past — whom does that officer report to, whose laws is the officer enforcing, whose budget pays the salary? The marriage license, the driver's license, the birth certificate of your child, the deed to your house, the building permit for your deck, the property tax bill, the parking ticket, the sales tax on your coffee?

Almost none of those things involves the federal government. Roads (mostly state and local), water (local), schools (local with state funding), police (local with sheriff at county), marriage licenses (county), driver's licenses (state), birth certificates (state), deeds (county), building permits (city), property taxes (local), parking tickets (city), sales tax (state and local). The federal government does enormous things — Social Security, Medicare, defense, monetary policy, civil rights enforcement, interstate commerce regulation, immigration. But the state-and-local layer is where most of the texture of governance touches your life.

This chapter does not ignore that asymmetry. It centers it. The empirical claim that drives the chapter is straightforward: by employees, by spending on services people actually use, and by sheer institutional volume, state and local government is the larger half of American government. Not the more glamorous half. Just the bigger one.

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

1. Describe the empirical scale of state and local government in the United States — how many units, how many employees, how much money.
2. Compare structures across the 50 states — bicameral and unicameral legislatures, strong and weak governors, methods of judicial selection, term limits.
3. Distinguish the units of local government — counties, municipalities, townships, school districts, special districts — and explain what each does.
4. Explain direct democracy mechanisms — initiative, referendum, recall — and analyze their effects.
5. Account for tribal and territorial governments as part of the American federal system, not as footnotes.
6. Apply the framework of asymmetric federalism from Chapter 4 to the policy areas where states diverge most dramatically — abortion, guns, drugs, criminal justice, education.
7. Locate yourself in this system — your state legislator, school board, county commissioner — as a Democracy Audit exercise.

I. The empirical picture

The 50 states + DC + 5 territories

There are 50 states. There is also the District of Columbia, which is not a state but has a population larger than two states (Wyoming and Vermont) and pays federal taxes without voting representation in Congress. DC residents vote for President — a right granted by the Twenty-Third Amendment in 1961 — and elect a non-voting delegate to the House. The push for DC statehood is one of the live constitutional questions of the era; it cuts on partisan lines because DC reliably votes Democratic and admission would add two reliably Democratic senators.

There are also five inhabited US territories: Puerto Rico (population ~3.2 million as of 2026, larger than 22 states), Guam (~170,000), the US Virgin Islands (~100,000), the Commonwealth of the Northern Mariana Islands (~50,000), and American Samoa (~50,000). Residents of these territories are generally US citizens (with American Samoa as a complicated exception: residents are US "nationals" but not citizens by birth). They pay some federal taxes, are subject to federal law, and serve in the US military at high rates. They do not vote for President, send only non-voting delegates to the House, and have no Senate representation. We will return to territorial status later in the chapter; it is one of the genuinely unresolved questions in the constitutional structure.

The 90,000 units of local government

Counting local-government units is a job of the US Census Bureau. The most recent Census of Governments (2022) found just over 90,000 distinct units of local government in the United States. The breakdown is approximately:

• Counties (or county-equivalents): about 3,000. Louisiana calls them "parishes," Alaska calls them "boroughs," and a few states (Connecticut, Rhode Island) have effectively eliminated county government. Some independent cities (St. Louis, Baltimore) are not in any county.
• Municipalities: about 36,000. Cities, towns, villages, boroughs (in some states). The legal distinctions among "city," "town," and "village" vary enormously by state.
• Townships: about 12,800, concentrated in the Midwest, the Northeast, and parts of the Mid-Atlantic. Most western states have no townships at all.
• School districts (independent of municipal government): about 12,800. Where most education policy actually happens. Some states (Hawaii, with one statewide district) are unusual.
• Special districts: about 38,000. Water, fire, library, sewer, transit, mosquito-abatement, parks, hospitals, soil conservation, ports. Often invisible to the citizens they serve.

That is around 90,000 governments — each with its own elected or appointed leaders, its own taxing or fee-charging authority, and its own narrow slice of public function. The federal government, by comparison, is a single government.

The 20 million employees

State and local governments together employ around 20 million people — roughly 5 million state employees and 15 million local employees. The federal civilian workforce, by comparison, is about 2.1 million (plus around 1.3 million active-duty military). Even counting active-duty military as federal "employees," state-and-local outnumber federal by about five to one.

Where do the employees go? The largest single category at the local level is public-school teachers (around 3.5 million). Then police and corrections officers (around 1.2 million). Then a long tail of nurses, social workers, road crews, librarians, water-treatment workers, fire fighters, sanitation workers, court clerks, parks staff, building inspectors, planners.

The spending

Federal spending in fiscal year 2025 was approximately $7.0 trillion. State and local combined spending was approximately $4.1 trillion. So the federal government still spends more — but the federal total includes Social Security, Medicare, defense, and interest on the national debt, which together account for over half of federal outlays. On the services people actually consume directly — schools, police, roads, prisons, courts, health and hospitals, sanitation, public housing, libraries, parks — state and local governments substantially outspend the federal government.

Where the money comes from

State governments rely on a mix of income taxes (44 states have one; the holdouts are AK, FL, NV, NH partial, SD, TN, TX, WA, WY), sales taxes (45 states have one; the holdouts are AK, DE, MT, NH, OR), corporate taxes, severance taxes (in resource states), federal grants, and fees. Local governments rely heavily on property taxes — by far the largest local revenue source nationally — plus sales taxes (in some states), fees, and intergovernmental transfers from state and federal sources.

The reliance on property tax has been a source of recurring fiscal pressure. Property is geographically immobile, which makes the tax stable, but the tax base is also geographically concentrated, which means schools funded primarily by property tax produce funding gaps between rich and poor districts. State equalization funds, federal Title I dollars, and a long line of state-court decisions on school-funding equity (Serrano in California, Robinson in New Jersey, Edgewood in Texas, DeRolph in Ohio) have reshaped local finance, but funding disparities remain a recurring issue.

II. State government structure

Three branches, with variations

All 50 state constitutions establish three branches: a legislature, an executive headed by a governor, and a judiciary. Most state constitutions include a bill of rights — and many of those bills of rights are broader than the federal Bill of Rights. The Massachusetts Declaration of Rights, dating from 1780 and drafted largely by John Adams, predates the federal Bill of Rights and contains rights provisions that the Massachusetts Supreme Judicial Court has read as protecting same-sex marriage (Goodridge, 2003) and a right to public education (McDuffy, 1993).

This is a feature of the system, not a bug. State constitutions establish the floor for federal rights — states cannot grant less than the federal Constitution — but they can grant more. This is especially significant after the Dobbs v. Jackson Women's Health Organization decision in 2022 returned abortion regulation to the states; in several states, abortion rights protections now flow from state constitutional provisions (often equal-protection or privacy clauses) read by state supreme courts.

State constitutions are also much longer than the federal one. The US Constitution is about 7,600 words. The average state constitution is about 39,000 words. Alabama's constitution, at over 350,000 words, is the longest constitution in the world — longer than the US Constitution by a factor of nearly 50. The reason is amendment patterns. Most state constitutions are far easier to amend than the federal one — by initiative, by majority vote in the legislature plus referendum, by simple majorities — and over time, the amendments accumulate. Many state constitutions read like a mix of fundamental law and ordinary statutory detail.

Bicameral, with one outlier

Forty-nine states have bicameral legislatures — a House (or Assembly, or House of Delegates) and a Senate. The structure echoes the federal model, though the rationales differ: state senators rarely represent fixed equal subdivisions the way US senators do, since Reynolds v. Sims (1964) requires both chambers of state legislatures to be apportioned by population.

The single outlier is Nebraska. Since 1937, Nebraska has had a unicameral, nonpartisan legislature of 49 senators. The structure was adopted on the urging of Senator George Norris (a Republican-progressive), who argued that bicameralism encouraged conference-committee deal-making in secret and that party labels at the state level were captured by national politics. The Nebraska legislature is officially nonpartisan: candidates appear on the ballot without party labels, and committee chairs are chosen by colleagues rather than by caucus. In practice party identifications are well-known and partisanship affects voting, but the formal nonpartisan structure shapes legislative behavior in measurable ways.

Nebraska is a genuine outlier. No other state has copied it. The fact that no other state has copied it does not by itself prove that bicameralism is better; it proves that bicameralism is the path of least resistance once it has been chosen. The Nebraska experiment is one of the live arguments in state-government theory.

Term limits

Fifteen states impose term limits on their state legislators. The list, as of 2026: Arizona, Arkansas, California, Colorado, Florida, Louisiana, Maine, Michigan, Missouri, Montana, Nebraska, Nevada, Ohio, Oklahoma, South Dakota. Most term-limited states adopted the limits via ballot initiative during the term-limits wave of the 1990s. Several states (Idaho, Utah, Massachusetts, Oregon, Washington, Wyoming) adopted limits and later repealed them, sometimes through legislative override and sometimes through ballot initiative.

The empirical research on legislative term limits is mixed but converges on a few findings. Term limits reduce institutional memory — newer legislators rely more heavily on staff, interest-group experts, and lobbyists for substantive guidance. Term limits increase legislative turnover and demographic diversity — more women and racial minorities have entered legislatures with term limits than without. Term limits shift power to the executive and the bureaucracy — when legislators leave just as they have learned how to legislate, governors and agency heads with longer institutional memories accumulate the comparative advantage. Term limits do not reduce campaign spending — candidates spend roughly as much, sometimes more, in open seats. Term limits do not reduce polarization — they may increase it slightly.

These findings are politically uncomfortable for both sides. Term-limit advocates wanted to reduce career politicians and the entrenchment of incumbency; the data suggests they accomplished part of that but at the cost of legislative competence. Term-limit opponents wanted to preserve experienced legislators; the data suggests they were right about the institutional cost but wrong that career politicians are reliably more responsive than newcomers. Reasonable people disagree about whether the trade is worth it.

III. Governors

Roles and powers

A governor is, simultaneously, the chief executive of the state, the head of state ceremonially, the leader of the state's political party, and the commander of the state's National Guard (subject to federal mobilization). The set of formal powers varies a lot across states.

Veto authority. Every governor has a veto. Forty-three states give the governor a line-item veto — the ability to strike specific items in an appropriations bill while signing the rest. The federal president has no line-item veto (the Clinton v. City of New York decision of 1998 struck down the federal version as a violation of bicameralism and presentment). State governors, operating under state constitutions, do.

Wisconsin's partial veto is the most expansive in the country. The Wisconsin governor can strike not only line items but specific words and even individual letters from a bill, in some cases changing the bill's meaning. (Earlier "Frankenstein veto" practices, where governors stitched together the surviving text into new legal meanings, were curtailed by amendment in 1990 and 2008, but the partial veto remains broader than any other state's.) Both Republican and Democratic Wisconsin governors have used it aggressively.

Pardon power. Most governors hold the pardon power for state convictions, sometimes in conjunction with a board of pardons. Some states (Connecticut, Georgia, Idaho, Nebraska, South Carolina, Utah) vest the pardon power in a board rather than the governor alone.

Executive orders. Governors issue executive orders affecting state agencies — much like the federal president — and the legal limits depend on the state constitution and statute. The COVID-19 era stretched gubernatorial executive-order practice in every state, with significant subsequent litigation in many states (notably Michigan, where the state supreme court limited the governor's emergency powers in 2020) about how far that authority extends.

Appointment power. Governors appoint cabinet officers, agency heads, judges in some states, and members of state boards and commissions. Senate confirmation is required in most states for the most significant appointments.

National Guard command. Governors command the state's National Guard except when federal authorities call the Guard into federal service. The federalization question is a recurring flashpoint — most recently in deployments related to the southern border, civil unrest in 2020, and the response to Hurricane Helene in 2024.

Strong governor vs. weak governor traditions

Some state constitutions concentrate power in the governor; others diffuse it across elected officials. The strong governor tradition is most associated with New York (where the governor controls the budget process and has strong veto and appointment powers) and California (where the governor is also a national figure with national implications for Democratic Party strategy).

The weak governor tradition is most associated with Texas. The Texas governor's formal powers are surprisingly limited. The Lieutenant Governor (separately elected) presides over the state senate and controls committee assignments, giving the office substantial agenda power. The Texas Comptroller, Attorney General, Land Commissioner, Agriculture Commissioner, and Railroad Commissioners are all independently elected. The Legislature meets only every other year (in regular session). The result: Texas governors operate in an institutional environment where formal powers are modest, and effectiveness depends heavily on personal political capital, alignment with the Lt. Governor, and use of the bully pulpit. (Greg Abbott, governor since 2015, has tested the limits of executive authority on border policy and other issues, with frequent litigation.)

Most states fall in between these poles. The National Governors Association and various political scientists have produced "Index of Formal Powers" rankings; New York, Massachusetts, and Maryland tend to score high, while Texas, Vermont, and Rhode Island tend to score lower.

Term limits on governors

Thirty-six states limit governors to two terms (consecutive or lifetime, varying). Fourteen states do not — including Connecticut, Idaho, Illinois, Iowa, Massachusetts, Minnesota, New Hampshire, New York, North Dakota, Texas, Utah, Vermont, Washington, and Wisconsin. Long-serving governors are mostly in this second group: Iowa's Terry Branstad (24 years total across two stints), New York's Mario Cuomo (12 years), Texas's Rick Perry (14 years).

Recent governors

The current generation of governors — as of 2026 — includes a number of nationally prominent figures whose styles illuminate the variety of the role:

• Gavin Newsom (D-California, 2019–): tested the limits of state executive power on COVID, climate, and homelessness; survived a recall in 2021; has been mentioned for higher office.
• Ron DeSantis (R-Florida, 2019–): pursued a high-profile conservative-populist agenda on education, public health, and immigration; ran for President in 2024.
• Greg Abbott (R-Texas, 2015–): ran on a platform emphasizing border security and state preemption of local policy.
• Kathy Hochul (D-New York, 2021–): assumed the office on Andrew Cuomo's resignation; won her own term in 2022.
• Gretchen Whitmer (D-Michigan, 2019–): a kidnapping target during COVID; survived to serve out her terms; mentioned for national office.
• JB Pritzker (D-Illinois, 2019–): leveraged a substantial personal fortune; positioned Illinois as a destination state on reproductive rights post-Dobbs.
• Chris Sununu (R-New Hampshire, 2017–): defines a moderate-Republican-governor lane in a competitive state.
• Jared Polis (D-Colorado, 2019–): the first openly gay person elected governor in the United States.
• Janet Mills (D-Maine, 2019–): the first woman to serve as governor of Maine.
• Spencer Cox (R-Utah, 2021–): cofounder of "Disagree Better," a bipartisan governors' initiative on civil discourse.

The list is illustrative, not exhaustive. The point is that governors are now national figures in ways that change the politics of the office. Modern governors run for President from a national platform built largely on what they did at home. The incentive structure shapes how the job is performed.

IV. State legislatures

Part-time, full-time, and hybrid

The National Conference of State Legislatures (NCSL) classifies state legislatures into three categories.

Full-time professional legislatures (often called "Red" legislatures in NCSL's color coding, though the terminology has nothing to do with party). These are large, full-time, well-staffed, well-paid bodies modeled on the US Congress. The clear examples: California, New York, Pennsylvania, Michigan, Ohio, Massachusetts, Wisconsin, Illinois, New Jersey. California state senators earn ~$130,000/year; New York legislators earn ~$142,000/year. They meet year-round or close to it.

Part-time citizen legislatures ("Blue" in NCSL's classification, again without partisan meaning). Small, low-paid, short-session bodies on the model of the original American legislatures. New Hampshire, Vermont, North Dakota, South Dakota, Wyoming, Montana, Maine. New Hampshire pays its legislators **$200 for the entire two-year term**. Wyoming pays $150/day during session. The expectation is that legislators have other jobs.

Hybrid ("White" in NCSL's classification). Most states fall here. Sessions are limited (typically 45–120 days per year or every other year), but legislators are paid a meaningful (though not full-time) salary, and there is professional staff. Texas, Florida, Georgia, Virginia, Washington, Colorado, and most of the rest are hybrid.

The differences matter. Full-time legislatures have substantial staff capacity, can take up complex legislation in detail, and develop substantive expertise across multiple sessions. Part-time legislatures rely heavily on the executive branch, on outside experts, and (critically) on lobbyists and interest groups for substantive guidance. The relationship between legislative professionalism and lobbyist influence has been studied extensively; the consensus is that less-professionalized legislatures are more susceptible to interest-group capture, though the effect varies by issue area.

The trade-off is also genuine. Full-time legislatures are more expensive, more polarized along national lines, and arguably less connected to ordinary life. Part-time legislatures are cheaper, more grounded in non-political work experience, and more limited in what they can accomplish in a 60-day session. Many state-government reformers have argued for one or the other; both arguments have legitimate empirical support.

Sessions

Sessions vary from continuous (California, New York) through limited (most states) to biennial (Texas, Nevada, Montana, North Dakota — four states whose legislatures meet only every other year, in regular session). The biennial pattern is a relic of the era when long-distance travel was hard and citizen-legislators returned home for the off year. It persists today partly because the underlying argument — less government is better government — retains political appeal in those states.

In Texas, the legislature meets for 140 days every other year. Anything not addressed in those 140 days waits until the next regular session, unless the governor calls a special session (which the governor controls absolutely — the legislature cannot extend or adjourn without consent in a special session). The agenda-setting power that this gives the governor is significant.

Polarization at the state level

The conventional wisdom holds that state legislatures, being smaller and more local, are less polarized than Congress. The data does not consistently support this. Several state legislatures — Wisconsin, Michigan, North Carolina, Pennsylvania, Arizona — score higher on standard polarization measures than Congress in some recent sessions, partly because state-legislative caucuses are more homogeneous (no need to keep moderate members happy in a 50-50 chamber when the chamber is 60-40).

Other state legislatures — Massachusetts, Maryland, Wyoming, Utah, where one party dominates — show lower surface polarization simply because the minority party is too small to matter. Within the dominant party, however, the major political conflicts often play out on factional lines (moderate Democrats vs. progressive Democrats in Massachusetts; conservative-establishment Republicans vs. populist Republicans in Utah).

The takeaway: state-legislative politics is its own thing. National partisan dynamics increasingly drive state-level voting patterns, but the local mix of dominant-party factions, regional interests, and state-specific issues still matters substantially.

V. State courts

Where the litigation is

Roughly 95% of all civil and criminal cases in the United States are filed in state courts, not federal courts. The state-court system is where most of American adjudication happens — divorce, child custody, traffic, small claims, eviction, contract, tort, the vast majority of criminal prosecutions, probate, family law, juvenile justice, mental-health commitments. A typical American who interacts with a court interacts with a state court.

Each state has its own court system. Most have a trial-court level (sometimes called "superior," "circuit," "district," or "court of common pleas"), an intermediate appellate court (in 41 states), and a state supreme court (which in some states, confusingly, is the trial court — New York's "Supreme Court" is the trial court; the appellate court is the "Court of Appeals"). The structures vary but the basic logic is similar.

Selection methods

How state-court judges are selected varies enormously and has been a subject of recurring reform debate.

Gubernatorial appointment with confirmation by a legislative body or council. New Jersey, Massachusetts, New Hampshire, Maine, Hawaii, Connecticut, Delaware, Rhode Island, Virginia (legislative election), Vermont, South Carolina (legislative election), and a few others. The federal model.

Nonpartisan election. Wisconsin, Minnesota, Michigan, Idaho, Oregon, Washington, Montana, Nevada, North Dakota, Kentucky, Arkansas, Mississippi (after 1994 reform), Georgia (after 1985), and more. Judges run on the ballot without party labels.

Partisan election. Texas (full partisan), Alabama, Illinois (initial selection), Louisiana, Pennsylvania, North Carolina (after 2018 reversion), West Virginia (until 2015 reform). Judges run with party labels.

Missouri Plan / merit selection with retention election. The governor appoints from a list provided by a nominating commission; the appointee runs unopposed in a "yes/no" retention election after a probationary term. Missouri (where the model originated in 1940), Alaska, Arizona, Colorado, Florida (appellate only), Iowa, Kansas, Nebraska, Tennessee, Wyoming, and others.

Combinations. Many states use different methods at different levels — for example, gubernatorial appointment for trial judges and election for appellate judges, or the reverse. New Mexico, Indiana, Tennessee, and others have hybrid systems.

The arguments are genuine on every side. Election advocates argue that judges who exercise enormous power over citizens' lives should be accountable to voters. Appointment advocates argue that judging requires legal expertise that voters cannot evaluate, that election turns judges into politicians (raising money, taking positions), and that judicial independence requires insulation from popular passions. Merit-selection advocates argue that the Missouri Plan combines the legitimacy of accountability (the retention election) with the protection of expertise (the nominating commission) — though critics argue the nominating commissions are themselves political and that retention elections are nearly always uncontested.

The empirical research is mixed but substantial. Judges in elective systems do, in fact, behave more politically — more sentencing variation tied to electoral cycles, more responsiveness to public opinion in death-penalty cases, more campaign donations from law firms with cases pending. Judges in appointive systems are more insulated but also less demographically representative. Judges in merit-selection systems trend toward the appointive end on most measures.

State supreme courts as constitutional actors

State supreme courts have always interpreted state constitutions, but their importance has grown dramatically in the post-Dobbs era. When Dobbs v. Jackson Women's Health Organization (2022) overturned Roe v. Wade and returned abortion regulation to the states, abortion-rights protections in many states became questions of state constitutional law. State supreme courts in Kansas, Montana, Florida, Michigan, and others issued decisions on whether state constitutions protect abortion access.

Other state-court constitutional doctrines that go beyond the federal floor:

• Massachusetts — Goodridge v. Department of Public Health (2003), reading the state Equal Rights Amendment to require recognition of same-sex marriage twelve years before Obergefell.
• California — Article I, Section 7 read more expansively than the federal Equal Protection Clause; state courts have extended affirmative-action and education-equity doctrines beyond the federal floor in past decades.
• New Jersey — Mount Laurel doctrine (since 1975) requiring municipalities to provide their fair share of affordable housing, an interpretation of state constitutional equality and welfare provisions with no federal analogue.
• Vermont — Baker v. Vermont (1999), reading the state's Common Benefits Clause to require recognition of same-sex relationships (initially through civil unions; full marriage came later).

The pattern is what scholars call the "new judicial federalism": state courts using state constitutions to provide rights protections beyond the federal floor. The doctrine cuts in multiple directions politically. Some state constitutions have been read to require additional rights protections; others have been amended to limit state-court interpretation, as in Florida's amendments restricting state-court authority over redistricting.

VI. Initiative, referendum, and recall

The mechanisms

The Progressive movement of the early 20th century pushed for direct democracy at the state level, on the theory that legislatures captured by railroads, banks, and other entrenched interests needed an end-run for popular policy. Twenty-four states now allow some form of citizen-initiated ballot measure.

Three distinct mechanisms:

Initiative. Citizens gather signatures to place a measure directly on the ballot. The measure becomes law if it passes. Some states allow initiatives to amend the state constitution; some allow only statutory initiatives; some allow both. California, Oregon, Colorado, Arizona, Florida, Massachusetts, Michigan, Missouri, Ohio, Washington — these are heavy-initiative states. Most southern, plains, and northeastern states do not allow initiatives; they require legislative action to put a measure on the ballot.

Referendum. A measure passed by the legislature is submitted to the voters for approval. Some referenda are mandatory (constitutional amendments require voter approval in nearly all states); some are optional (the legislature chooses to ask the voters); some are triggered by petition drive (citizens force a legislative measure to a vote, called a "popular referendum" in some states).

Recall. Voters may remove an elected official before the term ends through a petition-and-vote procedure. Eighteen states allow recall of state officials; many more allow recall of local officials.

California, the heaviest user

California is the heaviest-user state for direct democracy and the closest case study. The state has had ballot initiatives for over a century. Notable examples include Proposition 13 (1978), capping property taxes and reshaping California's fiscal structure for decades; Proposition 187 (1994), restricting public services to undocumented immigrants (struck down by federal courts before full implementation); Proposition 8 (2008), defining marriage as between a man and a woman (later struck down by federal courts); and a long series of measures on tax policy, criminal justice (the three-strikes law and its later modifications), and now (in the 2020s) housing and labor regulation.

Critics of California's heavy reliance on direct democracy point to several costs. The state budget is constrained by initiative-imposed spending requirements (Proposition 98, on K-12 education) and revenue restrictions (Proposition 13). The legislature, in some accounts, has been weakened by the easy availability of going around it. Well-funded interest groups can dominate the signature-gathering process; the cost of qualifying for the California ballot now runs into the millions. Complex tax and regulatory questions are reduced to up-or-down votes, with summaries written by people who have stakes in the outcomes.

Defenders argue that direct democracy has produced policy that the legislature would never pass because of interest-group capture, that voters in the aggregate make defensible decisions on most measures, and that the device gives citizens a meaningful voice between elections. Both arguments have merit; reasonable people disagree about whether the trade is worth it.

TABOR and the Colorado experience

Colorado's Taxpayer's Bill of Rights (TABOR), passed by initiative in 1992, requires voter approval for tax increases and limits revenue growth to inflation plus population growth. Surplus revenue is refunded to taxpayers. TABOR has been continuously controversial — credited by supporters with restraining government growth, blamed by critics for underfunded schools and infrastructure. Subsequent ballot measures (notably Referendum C in 2005) have suspended portions of TABOR. The Colorado experience is the clearest test case for fiscal limitations imposed by direct democracy.

Recent abortion-rights initiatives

In the post-Dobbs era, ballot measures on abortion rights have become a central feature of state-level politics:

• Kansas (August 2022) — voters defeated a constitutional amendment that would have removed abortion-rights protection from the state constitution. The vote was 59% no, 41% yes, in a state Trump won by 14 points in 2020.
• Michigan (November 2022) — voters approved Proposal 3, a constitutional amendment protecting reproductive rights. 57% yes.
• Vermont (November 2022) — voters approved a similar amendment. 77% yes.
• California (November 2022) — voters approved Proposition 1, also protecting reproductive rights. 67% yes.
• Kentucky (November 2022) — voters defeated a constitutional amendment that would have specified no right to abortion in the state constitution. The amendment lost 52-48 in a state Trump won by 26 points.
• Ohio (November 2023) — voters approved Issue 1, a constitutional amendment on reproductive rights. 57% yes.
• Florida (November 2024) — a constitutional amendment protecting abortion rights received 57% yes, but Florida's threshold for constitutional amendments is 60%, so the amendment failed despite majority support.
• Missouri (November 2024) — voters approved Amendment 3, protecting abortion rights. 52% yes.
• Arizona (November 2024) — voters approved Proposition 139. 62% yes.

The pattern is empirically clear: in every state with a direct-democracy mechanism that has voted on abortion rights post-Dobbs, the pro-abortion-rights position has won, often by substantial margins, often in states that vote Republican for federal office. The pattern complicates simple narratives about either party's ownership of "popular opinion" on the issue. Both parties have absorbed the lesson; the strategic implications are still being worked out.

Recall

Recall elections are rare but consequential when they occur. Gray Davis (D-California, 2003) was recalled and replaced by Arnold Schwarzenegger (R), the only successful gubernatorial recall in over 100 years (until North Dakota's Lynn Frazier in 1921, the most recent precedent before Davis). Scott Walker (R-Wisconsin, 2012) was the target of a recall after his collective-bargaining legislation; he survived, becoming the first US governor to defeat a recall. Gavin Newsom (D-California, 2021) was the target of a recall during the COVID-19 era; he survived with a 62-38 margin. The threshold for triggering a recall (signature thresholds, allowed grounds) varies enormously by state.

VII. County government

What counties do

Roughly 3,000 counties (or county-equivalents) cover approximately the entire United States outside of a few independent cities. Counties are the geographic-administrative middle layer between states and municipalities. They typically run:

• The sheriff's office, providing law enforcement in unincorporated areas and operating the county jail.
• The county courts (sometimes the trial-court level for the state).
• Property records — deeds, mortgages, liens, the public title chain.
• Property tax assessment and collection (sometimes shared with municipalities).
• Public health departments, which became national news during COVID-19.
• Vital records (birth certificates, death certificates, marriage licenses).
• Many road and bridge systems, especially in unincorporated areas.
• Voter registration and election administration in many states.

There are about 3,000 sheriffs in the United States, each typically elected at the county level. They command around 300,000 sworn deputies and corrections staff. The American sheriff is a constitutional officer in many states, accountable directly to county voters rather than to the governor or state attorney general. This produces real friction in some states — notably California (where some sheriffs have refused to enforce certain state firearm or sanctuary policies) and Texas (where some sheriffs have aligned closely with state border-enforcement priorities). Sheriff politics has become more nationally visible in recent years.

County size variation

The variation in county size and capacity is enormous. Los Angeles County, California has a population of approximately 9.7 million — larger than 41 states. Its annual budget exceeds $40 billion. Its Department of Health Services is one of the largest hospital systems in the country. The Los Angeles Sheriff's Department has about 18,000 employees.

At the other extreme, Loving County, Texas has a population of about 50 people. Kalawao County, Hawaii (the former leprosy settlement on Molokai) has under 90. Wyoming and Nevada have multiple counties with under 1,000 residents.

The legal structures are roughly similar — elected boards, sheriffs, clerks, assessors — but the operational reality is wildly different. A sparsely-populated rural county may have one full-time sheriff's deputy and rely on the state highway patrol for backup; an urban county runs a full-scale municipal-law-enforcement operation.

VIII. Municipal government

Forms of city government

The structure of municipal government in the United States falls into roughly three forms.

Mayor-council government. A separately elected mayor leads the executive branch; a council serves as the legislature. This form comes in two variants:

• Strong-mayor systems give the mayor substantial executive power — appointment of department heads, veto power over council, control of the budget process. Major cities using strong-mayor: New York, Los Angeles, Chicago, Houston, Philadelphia, Boston, San Francisco, Atlanta, Detroit. The strong-mayor form is most common in large cities.
• Weak-mayor systems vest executive authority more diffusely — department heads may be elected separately or appointed by council, the mayor is largely a ceremonial head of council, and the council exercises substantial executive power. Some smaller and mid-sized cities still use this form.

Council-manager government. The council is elected; it hires a professional city manager to run day-to-day operations. The mayor is typically chosen by the council from among its members or elected directly but with limited formal power; the manager handles administration. This is the most common form among American cities of populations over 2,500 — over half of municipalities of that size use council-manager. Major cities using council-manager include Phoenix, San Antonio, Dallas, Charlotte, Cincinnati, Kansas City (MO), Sacramento, San Jose, Toledo. The form has Progressive-era origins; reformers viewed professional management as the antidote to political-machine corruption.

Commission government. A small group of commissioners is elected; each commissioner heads a city department (police commissioner, public-works commissioner, etc.). The form was popular in the early 20th century but has nearly disappeared — Portland, Oregon was the last major city to retain it, and Portland voters approved a switch to a different structure in 2022 (effective 2025).

What cities do

Cities run police (most US police officers work for municipal police departments, not county sheriffs or state police), fire (mostly municipal in cities; mostly county or special-district in rural areas), water and sewer, parks, garbage collection, building codes, zoning (with sometimes-momentous consequences for housing supply), local roads, public housing (in many cities), and the schools (in some — others have separate elected school boards independent of municipal government).

Land-use regulation may be the single most consequential thing many cities do. Zoning decisions made at the city level shape housing supply, neighborhood character, and economic geography. The American "missing middle housing" debate, the YIMBY/NIMBY arguments, the question of whether local zoning is sustainable in a high-housing-cost economy — all of these unfold mainly at city-council meetings, design review boards, and planning commissions.

Dillon's Rule and state preemption

Cities are creatures of state law. Dillon's Rule, articulated by Iowa Supreme Court Justice John Forrest Dillon in 1868, holds that municipal corporations have only those powers expressly granted by the state, those necessarily implied, and those essential to the declared objects of their charters — and any reasonable doubt is resolved against the city. Most states follow Dillon's Rule by default; some states (about 10) follow the looser "home rule" tradition, granting cities broader authority by constitutional or statutory delegation.

The implication is that state legislatures can preempt municipal action. In recent decades, state preemption has been one of the live arenas of state-and-local politics, and it cuts in both partisan directions:

• State preemption of local minimum-wage increases — Florida (which has been complicated by initiative-driven state-level minimum wage increases), Missouri, Wisconsin, and many others have prohibited cities from setting minimum wages above the state floor. Cities like St. Louis, Kansas City, and Birmingham have had local wage increases struck down by state preemption.
• State preemption of local gun ordinances — most states have preemption statutes prohibiting local firearm regulations that exceed state law. The pattern goes back to the 1970s in some states; recent examples include Pennsylvania, Florida, and Mississippi.
• State preemption of local sanctuary policies — Texas SB 4 (2017) and similar legislation in Florida and elsewhere preempted local jurisdictions from limiting cooperation with federal immigration enforcement.
• State preemption of local plastic-bag bans — Arizona, Texas, and several others.
• State preemption of local short-term-rental regulations — Florida, Tennessee, and others.
• Blue-state preemption running the other direction — Colorado and a few others have preempted some localities on rent-control, gun-shop bans, and other rules. New York's preemption of New York City's transportation policy on certain issues reflects the same dynamic.

The pattern is bipartisan. The party in control of the state preempts the party in control of certain cities, regardless of which party that is. Preemption is, in a sense, federalism's mirror at a different scale: just as the federal government preempts states on some questions, states preempt cities on others. The constitutional and political logic is the same, even if the values cut differently.

IX. School districts

The institutional facts

There are about 12,800 school districts in the United States, plus a smaller number of state-administered schools (Hawaii operates as a single statewide district). Most American K-12 students attend public schools governed by an elected school board at the local district level.

School-board members are usually elected on nonpartisan ballots in low-turnout elections, often held off-cycle from the major federal and state elections. Until recently, school-board elections were among the most under-attended in the country; turnout often ran in the single digits as a percentage of eligible voters.

This has changed substantially since 2020. School-board elections, after the COVID-era debates about reopening, mask mandates, curriculum, and book selection, have become centers of organized political contestation. Both ideological camps have invested in school-board races. National organizations — Moms for Liberty on one side, the National Education Association and various local advocacy groups on the other — have made school boards a strategic priority. Turnout has risen; spending has risen; the level of conflict at school-board meetings has risen.

Local property-tax funding

Most American school districts are funded primarily through local property taxes, supplemented by state aid (usually targeted at lower-wealth districts) and federal funds (most prominently Title I for lower-income students and IDEA for special education). Local property taxes typically fund about 35-45% of school spending; state aid provides 45-55%; federal funds about 8-12%. The mix varies enormously by state — Vermont (after Act 60) collects most school revenue at the state level; Connecticut and Massachusetts retain heavily-localized funding.

The reliance on property taxes produces inequality between districts. A district with a large commercial tax base or expensive houses can fund schools at much higher per-pupil rates than a low-income district at the same tax effort. State equalization formulas (some by state-court mandate after equity litigation) attempt to compensate, but funding gaps persist.

The empirical question of whether more spending produces better outcomes has a substantial literature. Recent meta-analyses (most prominently the work of Kirabo Jackson and colleagues) indicate that increased school spending does produce measurable gains in student outcomes — most clearly when targeted at lower-income districts and at specific evidence-based interventions. The effect is not automatic; spending alone does not guarantee outcomes. But the older finding (associated with the Coleman Report in 1966 and the Hanushek line of work in subsequent decades) that school spending shows weak effects on student outcomes has been revised in light of more recent research designs.

This is a contested but live empirical area. Reasonable people disagree about the policy implications.

X. Special districts

The 38,000

There are roughly 38,000 special districts — narrow-purpose government units created to handle a specific function. Water districts, fire districts, library districts, sewer districts, hospital districts, mosquito-abatement districts, transit districts, port districts, parks-and-recreation districts, soil-conservation districts, irrigation districts.

Most special districts have their own taxing authority — usually property tax, sometimes special assessments or fees — and their own elected or appointed boards. Most operate below the public's radar. Voter turnout in special-district elections (where the seats are elective) is typically very low.

The fiscal scale is substantial. Special-district spending nationally is over $200 billion per year. Some single districts run major operations: the Metropolitan Water District of Southern California serves 19 million people; the New York Metropolitan Transportation Authority runs subway, bus, and commuter-rail systems for the New York metropolitan area.

Why do special districts proliferate? Several reasons. They allow specialized expertise to govern technical functions (water-quality engineers running water districts rather than mayors). They allow geographic boundaries to match service areas (a fire district may serve parts of multiple municipalities). They allow taxing authority that bypasses general-government tax limits (in California, special districts have been used to fund services that Proposition 13 made hard to fund through general property tax). And they allow the political accountability for unpopular taxes to be distributed (a water-district board, not the city council, raises water rates).

The democratic-accountability problem is real, however. Citizens have limited ability to monitor 38,000 special districts. The number of elected positions across all special districts nationally runs into the hundreds of thousands. Most are filled in low-information elections. The combination of taxing authority and low public scrutiny is, by some accounts, a recurring source of governance failure (water-district scandals, hospital-district financial collapses) and, by others, a feature that has allowed American local government to specialize and adapt.

XI. Tribal governments

Sovereign nations

There are 574 federally recognized Native American tribes as of 2026. Tribal governments are sovereign nations within the United States, with a relationship to the federal government grounded in treaties, federal statutes, and a long line of Supreme Court decisions stretching back to Cherokee Nation v. Georgia (1831) and Worcester v. Georgia (1832).

Tribal sovereignty is genuine and complicated. Tribes have their own constitutions, courts, legislatures, executive officers, police forces, schools, health services, and economic-development authorities. They have substantial (though not unlimited) authority to regulate conduct on tribal land, levy taxes, and manage natural resources. They have sovereign immunity in most contexts, comparable to the sovereign immunity of states.

The relationship to states is layered and contested. Tribes are not part of state governments; tribal lands within a state are subject to a complex mixture of tribal, federal, and (in some areas) state jurisdiction. The doctrine that has emerged from the Court is that, in general, states do not have authority over tribal members on tribal land unless Congress has expressly granted it (Public Law 280, in the states it covers, is the major exception).

McGirt v. Oklahoma

The most consequential recent case on tribal sovereignty is McGirt v. Oklahoma (2020), in which the Supreme Court — in a 5-4 opinion by Justice Gorsuch — held that the Muscogee (Creek) Nation reservation in eastern Oklahoma had never been formally disestablished by Congress, and therefore much of eastern Oklahoma (including the city of Tulsa) is, for purposes of federal criminal law, "Indian country." The decision dramatically expanded the reach of tribal and federal jurisdiction over crimes committed on what is legally tribal land. Subsequent litigation has worked out the implications.

The decision was a striking reaffirmation of treaty obligations made by the United States to tribes in the 19th century. It also illustrates how federal-Indian law continues to be a live area of constitutional development, not a closed historical chapter.

What tribal governments do

Tribal governments run a wide range of public functions. Tribal courts hear civil and criminal matters under tribal law. Tribal police forces patrol tribal lands. Tribal departments of education run schools (often in collaboration with state and federal funding). Tribal health services, sometimes in coordination with the federal Indian Health Service, deliver care. Tribal economic-development authorities run gaming operations (where state-tribal compacts permit), other businesses, and resource-extraction ventures.

The economic role of tribal gaming has been substantial since the Indian Gaming Regulatory Act of 1988. Some tribes have used gaming revenue to fund extensive social services, schools, and member dividends; the largest tribal gaming operations (in California, Oklahoma, Connecticut, Florida) generate hundreds of millions in annual revenue. The economic uplift varies enormously by tribe; some tribes are wealthy by any standard, while others remain among the most economically distressed populations in the country.

Water rights under doctrines like Winters v. United States (1908) — the federal-reserved-water-rights doctrine — give some tribes substantial claims to water in arid Western states, with consequences working out in current litigation over the Colorado River, the Klamath, and other systems.

The chapter cannot do justice to the full complexity. The point is that tribal governments are part of the American federal system, not adjacent to it; they exercise sovereignty that other governments respect; and they deserve real treatment in any honest account of how American government works.

XII. Territorial governments

Five territories

The United States has five inhabited territories, each with a distinct history and governance arrangement.

Puerto Rico has been a US possession since 1898 and a "Commonwealth" since 1952, when its constitution was ratified. The island has an elected governor (the current governor is Jenniffer González-Colón, in office since 2025), a bicameral legislature, and a court system. Puerto Ricans are US citizens but cannot vote for President while resident in Puerto Rico. The non-voting Resident Commissioner represents Puerto Rico in the US House. Puerto Rico's status — statehood, independence, "enhanced commonwealth," current status — has been the subject of multiple plebiscites, none of which has produced a sustainable political path forward. The 2024 plebiscite was the most recent.

The fiscal crisis of the 2010s, the catastrophic effects of Hurricane Maria in 2017, and the federal oversight board (PROMESA) imposed in 2016 have all reshaped Puerto Rican politics. The population of the island (about 3.2 million) is larger than 22 states. The lack of full federal political rights is one of the live constitutional questions.

Guam has been a US territory since 1898. Population approximately 170,000. Substantial US military presence (Andersen Air Force Base, Naval Base Guam). Guamanians are US citizens; they cannot vote for President; a non-voting delegate represents Guam in the US House. The strategic importance of Guam in the Pacific has made it a focus of recent military buildup tied to Indo-Pacific strategy.

The US Virgin Islands has been US territory since 1917 (purchased from Denmark). Population approximately 100,000. Residents are US citizens; non-voting House delegate; no presidential vote. The territory's economy depends heavily on tourism and federal transfers; it has experienced repeated fiscal crises.

The Commonwealth of the Northern Mariana Islands (CNMI) entered the US relationship in 1986. Population approximately 50,000. Residents are US citizens; non-voting House delegate.

American Samoa is the unusual case. It has been a US territory since 1900. Population approximately 50,000. American Samoans are US nationals but not US citizens by birth (those born in American Samoa are not "born in the United States" for the Citizenship Clause; they are nationals — they can travel freely, work in the United States, and serve in the military, but they cannot vote in any election in any state without first naturalizing). The status has been the subject of repeated constitutional litigation; the US Court of Appeals for the DC Circuit ruled in 2021 that Congress, not the courts, must decide whether to extend birthright citizenship to American Samoa. American Samoa is unique in having retained its traditional fa'a Samoa governance structure (with the matai system), which the local territorial government recognizes alongside elected institutions.

The pattern

The five territories share several features. Residents pay some federal taxes (varying by territory and individual circumstances). Residents are subject to most federal laws. Residents serve in the US military at high rates — typically among the highest per-capita rates in the country. Residents do not vote for President. Their congressional representation is non-voting.

The collective population of the territories is roughly 3.5 million — comparable to four small states combined. The constitutional questions about their status have been before Congress for decades and have not been resolved. In the academic literature, scholars like José Cabranes, Christina Duffy Ponsa-Kraus, and Sam Erman have argued that the Insular Cases of 1901 — the Supreme Court decisions that established the territories' anomalous status — rest on doctrines that are inconsistent with the rest of constitutional law and that should be reconsidered. Whether and how this will happen is unclear.

The point for this chapter: when American Government textbooks ignore the territories, they are ignoring the political situation of about 3.5 million Americans. That is a choice. This book chooses otherwise.

XIII. Where state-and-local action diverges from federal

The chapter has presented the structures. The consequence of those structures, in the post-2010 environment of intensified partisanship and decentralization-by-court-doctrine, is that policy in many areas now varies dramatically by state. This is the asymmetric federalism point introduced in Chapter 4, made concrete.

Abortion access. After Dobbs, abortion regulation is a state question. As of 2026, abortion is broadly legal in roughly 25 states (often by state constitutional protection); banned at conception or near-conception in roughly 13 states; restricted at varying gestational ages elsewhere. The interstate landscape changes nearly every quarter as new litigation, ballot measures, and legislative action move the lines.

Gun laws. State firearm laws vary enormously. Some states require permits to purchase, registration, training requirements; others allow permitless concealed carry. The Bruen decision of 2022 has compressed the range of allowable regulation, but substantial variation remains. As of 2026, about 28 states have some form of permitless ("constitutional") concealed carry; about 22 states retain permit requirements.

Marijuana. As of 2026, recreational marijuana is legal in 24 states plus DC; medical use is legal in over 38 states; about 10 states still prohibit any recreational use. Federal law continues to classify marijuana as a Schedule I controlled substance, though the de facto enforcement landscape has shifted dramatically. The conflict between state and federal law is one of the genuinely odd areas of American legal practice.

Criminal justice. Sentencing, parole, capital punishment, juvenile justice, drug-policy enforcement, qualified-immunity reform — all are largely state matters and vary enormously. As of 2026, 23 states have abolished the death penalty (including five states that did so via judicial decision or legislative action since 2010). Texas, Florida, Alabama, Oklahoma, and Georgia continue to execute people; many states with death-penalty statutes on the books have not executed anyone in years.

Education curricula. Standards, content, book selection, sex education, history curricula — these are state and local. The post-2020 contests over CRT-related instruction, divisive concepts, library content, and gender-related curriculum have been state-by-state and district-by-district. Florida, Texas, and a number of other states have passed legislation restricting certain instructional content; California, Illinois, and others have moved in the other direction.

Election administration. Congressional and presidential elections in the United States are administered by states and counties. Voter ID requirements, mail-in voting rules, registration deadlines, drop-box policies, ballot-curing procedures, signature-matching standards — all are state-and-local. The variation is real.

The phrase "the laboratories of democracy," from Justice Brandeis, captures the upside — variation lets us see what works. The downside — what political scientist Robert Mickey has called "uneven democracy" — is that the variation can be stark, and the variation matters for citizens' lives. A pregnant person's options, a person carrying a firearm, a defendant in a capital case, a student in a public school: the federal Constitution defines the floor, but the practical reality is mostly state and local.

XIV. Closing: The Democracy Audit at the state-and-local level

The Democracy Audit project that runs through the book has been mostly federal so far — your representative in Congress, your senators, your presidential vote. This chapter asks you to extend it.

Find your state legislator — both the state-house member and the state-senator. Look up their committee assignments, their bill sponsorships, their votes on the bills you care about. State legislators are often more accessible than federal representatives. They take constituent calls. They show up at town halls. They will sometimes send you a personal email response.

Find your county commissioner (or board of supervisors, or freeholder, or whatever your state calls the county legislative body). Look up the most recent county budget. Note where the money goes. Note the largest single line items. (For most counties, it is the sheriff's office and the schools — though the schools may be funded through a separate district.)

Find your city council district if you live in a city. Look up the most recent council agenda. Note the kinds of decisions being made — zoning variances, business licenses, ordinance changes. These are the decisions that shape what your neighborhood looks like.

Find your school board. Read the agenda for the next meeting. Note the kinds of items on the agenda — staffing, curriculum adoption, budget approval, parent comment.

Find a special district that touches you. The water you drink, the library card in your wallet, the bus you ride, the park you walk in, the fire engines that pass — at least one of these is run by a special district. Look up the board and the budget.

If you find that the state legislator is uncontested, the school-board meeting is sparsely attended, the special district is largely invisible — note that. Note how much governance happens with very little public scrutiny. Note the leverage that a few engaged citizens can have at this level.

This is not a moral exhortation that everyone must become a school-board volunteer. It is a piece of empirical political analysis. State and local government is where most government actually happens. It is, on average, less reported, less polarized, less staffed, and more accessible than the federal government. Citizens who care about how government works have, at this level, a higher marginal-return-on-effort than at any other.

That is, in some sense, the recurring lesson of this whole book: power flows to those who show up. At the state-and-local level, "showing up" can mean a single email to a school-board member, a single comment at a planning commission, a single visit to your state representative's local office. The leverage is real.

The federal government will continue to dominate the news cycle. The state-and-local level will continue to dominate your daily life. The asymmetry between attention and consequence, in this domain, runs in a direction worth noticing.

Chapter 16 turns to political parties — how they organize, how they nominate candidates, how they raise money, how they have changed over American history, and what role they now play in connecting citizens to the institutions this section has described.