Chapter 19: Political Parties — The Two-Party System, Realignment, and Why Third Parties Can't Win

19.1 The Things We Call "the Parties"

Walk into any American conversation about politics and the words "the Republicans" or "the Democrats" will appear within five sentences. Almost no one stops to ask what those words refer to.

They refer, depending on context, to at least three very different things.

When a person says "I'm a Republican" or "I'm a Democrat," they are usually talking about themselves — their identity, their team, the box they would check on a survey. The party in this sense is a label that lives inside the head of an individual voter. Political scientists call this party identification, and it has turned out, since the 1950s, to be one of the most stable and powerful predictors of political behavior. People change religions more often than they change parties.¹

When a journalist writes that "the Democrats blocked the bill," they are usually talking about the party in government — the senators, representatives, governors, and other officeholders who share the label. The Democrats in this sense are 47 senators and 213 House members and a Cabinet (when they hold the White House) and judicial nominees and committee staff and a coordinated leadership team that decides how the team will vote. The party in government is what produces legislation, oversight, and confirmations.

When the same journalist writes that "the Republican National Committee announced," she is talking about the party as organization — the formal entity, with bylaws, a chair, a budget, a building in Washington, state-level affiliates, county-level committees, and a set of rules for selecting nominees and conducting party business. The RNC has a chair, a treasurer, members from every state, and an annual budget in the hundreds of millions of dollars. The DNC has the same. The state parties — fifty Republican state parties and fifty Democratic state parties, plus territorial affiliates — vary widely in capacity and influence. Local parties, once the engine of the urban political machines that dominated American cities for the first half of the twentieth century, are mostly thin, volunteer-run operations today.

The party-as-label, the party-as-government, and the party-as-organization are connected but not identical. A party-as-organization can be weak (the RNC and DNC have substantially less control over their nominations than the formal organizations of European parties) while the party-as-label is strong (party identification still predicts vote choice better than any single demographic variable). A party-as-government can be unified on most votes (party-line voting in Congress is at historically high levels) while the party-as-organization is internally divided (the 2016 Republican primary was a case of the organization losing control of who got to wear the label). A party-as-label can shift its underlying coalition (the Democratic Party of 1965 and the Democratic Party of 2025 share a name and not much else, in terms of which kinds of voters reliably vote for it) while the formal organization persists with continuous bylaws.

Throughout this chapter we will keep these three faces of a political party in mind. We will be specific about which face we are discussing whenever the distinction matters.

This chapter sits in the middle of Part III. Chapter 17 introduced public opinion — the raw material of mass politics. Chapter 18 examined the media — the channel through which most public opinion is shaped. The chapters that follow this one cover elections (20), campaign operations (21), voting behavior (22), and the way identity intersects with all of the above (23). Political parties are the institutions that connect public opinion, media, elections, and campaigns into something that can govern. They are also, in the United States, locked into a particular structural form — two and only two major parties, with brief and almost-always-failed third-party challenges around the edges. Why that structural form has held for nearly two centuries, and what the current parties look like inside that structure, are the questions this chapter is built to answer.

19.2 What Political Parties Are For

Before we examine the U.S. parties in detail, it helps to ask the foundational question: why do political parties exist at all? The Founders did not want them. George Washington's 1796 Farewell Address warned, with some specificity, against "the baneful effects of the spirit of party generally." The Constitution does not mention political parties anywhere. The original electoral system was designed assuming candidates would run as individuals on their own merits, with the second-place finisher becoming vice president. By 1800 — twelve years after ratification — the Founders had organized themselves into two competing factions, run a ferociously partisan election, and produced a constitutional crisis that required the Twelfth Amendment to fix. Parties emerged anyway. Why?

The deepest answer is that democracy at scale requires aggregation. A society of more than three hundred million people cannot make every collective decision by referendum. Voters cannot personally evaluate every potential officeholder for every position from school board to Supreme Court. They need shortcuts. A party label is a shortcut: if you know that a candidate is a Democrat, you can predict, with reasonable accuracy, how that candidate will vote on most major issues. If you know a candidate is a Republican, the same is true. The label saves the voter the immense cognitive cost of researching every candidate from scratch. Whatever its costs, party identification is information-economical. It is, in the political-science literature, the most efficient single piece of information a voter can have.²

Parties also do four things at the institutional level that no other entity does as effectively.

First, they recruit candidates. Someone has to find people willing to run for the 7,300 state legislative seats and the 535 congressional seats and the thousands of county and municipal offices that make up the American political system. Parties — through their state and local affiliates — do most of that recruiting. They identify potential candidates, vet them, encourage them to run, and connect them to donors, consultants, and volunteers.

Second, they coordinate campaigns and messaging. A party brand provides candidates with a ready-made platform, a recognizable label, and access to coordinated advertising, voter contact, and rapid response. Without parties, every candidate would have to build a campaign infrastructure from scratch. With parties, candidates inherit decades of accumulated organization.

Third, they aggregate diverse interests into governing coalitions. The Democratic Party of 2025 includes labor union members and college-educated suburbanites and African-American Protestants and secular progressives and a plurality of Hispanic and Asian voters. The Republican Party of 2025 includes evangelical Protestants and small-business owners and military families and working-class non-college whites and a growing share of younger Hispanic men. These groups do not agree on everything. They agree on enough to pull the same lever, and the party institutionally negotiates what the platform commitments to each component group will be. That negotiation — sometimes called "intra-party democracy" — is one of the messiest and most consequential features of the American system.

Fourth, they provide accountability. When voters are unhappy with how the country is being governed, they can throw out the party in power. They do not have to identify which specific senator failed them; they can vote against the team. The two-party system makes this dichotomous accountability simple in a way that multi-party coalition governments do not. Whether that simplicity is an advantage or a disadvantage is a question we will return to in §19.10.

The Founders' worry about parties was not, in retrospect, wrong. Parties do produce factional conflict; they do simplify complex policy disagreements into team identities; they do reward partisan loyalty over independent judgment. But the alternative — politics without parties — turns out, empirically, to be unsustainable in any modern democracy. Every functioning democracy on earth has political parties. The question is not whether to have them but how many, how organized, and how disciplined.

19.3 The Two-Party System: Why There Are Two and Only Two

The most important structural fact about American politics is that there are two major parties and approximately zero successful third parties. This has been true, with exceptions we will catalogue in §19.10, since the 1850s. It is a striking fact: most consolidated democracies have three or more parties that hold seats in the legislature and that participate in governing coalitions. The United Kingdom has a major-party system that occasionally tilts toward two parties but routinely seats Liberal Democrats, Scottish Nationalists, Greens, and others. Germany has a five- or six-party system. France oscillates between bipolar coalitions and multi-party fragmentation. Israel and the Netherlands routinely seat ten or more parties. The United States has, almost continuously since 1860, two and only two parties that hold the White House, a House majority, or a Senate majority.

The standard explanation is Duverger's Law, named for the French political scientist Maurice Duverger, who articulated the relationship in the 1950s.³ Duverger's Law in its simple form holds that single-member-district plurality voting tends to produce a two-party equilibrium. Each part of that phrase is doing work.

A single-member district is one in which a single seat is at stake — one House district elects one member; one state elects one senator (each Senate election is a single-seat contest); one nation elects one president (the Electoral College complicates this but does not change the underlying single-seat character). Compare to multi-member districts, in which a district might elect five or ten members, with the seats divided among parties in proportion to vote share. Most European democracies have at least some multi-member districts; the U.S. has essentially none above the school-board level.

Plurality voting means the candidate with the most votes wins, even if that is less than 50%. The U.S. uses plurality voting almost everywhere except in places that have adopted Ranked-Choice Voting (Maine for federal races, Alaska for federal and state races, several Bay Area cities, New York City for primaries) or runoff systems (Georgia's general-election runoff for U.S. Senate; Louisiana's "jungle primary"; Mississippi's runoff for state offices when no candidate clears 50%).

When you combine single-member districts with plurality voting, the result is a structural pressure toward two-party equilibrium. The mechanism has two parts:

First, the mechanical effect. In a single-member-district plurality system, only the largest plurality wins the seat. A party that consistently finishes third in most districts gets zero seats, even if it has 15% of the national vote. Third-party support is, in a structural sense, wasted — it does not translate into representation. Compare proportional representation: a party with 15% of the national vote in a PR system gets approximately 15% of the legislative seats, regardless of where its voters are concentrated. PR rewards small parties; plurality punishes them.

Second, the psychological effect. Voters who understand the mechanical effect anticipate it. If you are a voter whose first choice is a small-party candidate but whose second choice is a major-party candidate, and you know that a vote for your first choice is unlikely to elect anyone, you face an incentive to vote strategically — to abandon your first choice in favor of your second choice, in order to influence the outcome between the two viable candidates. This is sometimes called "strategic voting" or, less neutrally, "voting for the lesser evil." Over time, the psychological effect reinforces the mechanical effect: small-party support shrinks because voters abandon it, which makes small-party performance even worse, which causes more voters to abandon it. Equilibrium: two parties.

Duverger's Law is a tendency, not a certainty. It has exceptions — countries with single-member-district plurality voting that nonetheless have three or more parties, usually because of strong regional variation. Canada has had four-party national politics for decades despite first-past-the-post voting, partly because the Bloc Québécois is regionally concentrated in Quebec and the New Democratic Party has regional bases in Ontario and the West. The United Kingdom shows the same pattern with Scottish nationalists. But the United States, lacking strong regional minor-parties, has settled into the cleanest version of the Duverger equilibrium that any large democracy displays.

The structural pressure toward two parties is not the only thing that produces two-party dominance in the U.S., though it is the strongest. Other contributors:

Ballot-access laws in most states require third-party candidates to gather large numbers of signatures (often tens of thousands) to appear on the ballot, and to do so on tight deadlines. Major-party candidates are exempt or face much lower thresholds. The 2020 Libertarian and Green presidential candidates each had to mount fifty separate state ballot-access campaigns; the Republican and Democratic nominees got on the ballot automatically in every state. This is a real, consequential, often-overlooked barrier.

Primary system structure. Because both major parties hold open or semi-open primaries, factional movements that would, in a multi-party system, form their own party often fight for control of one of the two existing parties instead. The Tea Party did not become a "Tea Party Party"; it ran candidates in Republican primaries. The democratic-socialist movement around Bernie Sanders did not form a Democratic Socialist Party; Sanders ran in Democratic primaries. The MAGA movement did not become a separate party; it took over the Republican Party from inside. We will return to this pattern in §19.10.

Campaign-finance rules. Federal matching funds have historically been structured to favor major-party candidates. Public-financing rules in some states require demonstration of "viability" that a third party often cannot meet. And the practical reality is that donors who want to influence outcomes give to candidates who can win, which means major-party candidates almost without exception.

Debate-access rules. The Commission on Presidential Debates requires candidates to poll at 15% across an average of qualifying surveys to participate in general-election debates. No third-party candidate has cleared this bar since Ross Perot in 1992. Without debate access, third-party candidates cannot reach the audience they would need to grow their support, which keeps them below 15%, which keeps them out of the debates. The bar is, in effect, self-perpetuating.

The two-party system, in short, is the product of structural pressure (Duverger), institutional rules (ballot access, debates), and the strategic behavior of factional movements (which work inside the major parties rather than against them). Reformers can change the rules — and §19.11 examines what happens when they do — but the structural pressure remains until the underlying voting system changes.

19.4 First-Past-the-Post and Its Alternatives

A short technical detour, because the alternatives matter when we discuss reform in §19.11.

First-past-the-post (FPTP), also called "single-member plurality," is the system used in most U.S. elections. Each district elects one member; the candidate with the most votes wins; majority is not required. This is the system the structural pressures of Duverger operate on.

Ranked-Choice Voting (RCV), also called "instant-runoff voting," asks voters to rank candidates in order of preference. If no candidate has a majority of first-place votes, the candidate with the fewest first-place votes is eliminated and his or her voters' second-place choices are redistributed. The process repeats until one candidate has a majority. RCV is used for federal races in Maine (since 2018), for federal and state races in Alaska (since 2022), and for various municipal elections including New York City Democratic primaries, Minneapolis, San Francisco, and many smaller jurisdictions.

Approval voting asks voters to vote for as many candidates as they approve of, rather than ranking them. The candidate with the most approvals wins. Fargo, North Dakota (2018) and St. Louis, Missouri (2020) have adopted approval voting for some elections.

Top-two primary systems put all candidates from all parties on a single primary ballot, with the top two finishers (regardless of party) advancing to a general-election runoff. California adopted top-two primaries for state and congressional races in 2010; Washington State adopted a similar system in 2008; Alaska's 2020 reform installed a "top-four primary" with RCV in the general. Louisiana has used a unique "jungle primary" for state offices since 1975, where if any candidate gets a majority in the open primary, that candidate wins outright.

Proportional representation (PR) systems allocate legislative seats to parties in proportion to their share of the vote. Most European democracies use some form of PR. PR requires multi-member districts and, in pure form, eliminates the structural pressure toward two-party equilibrium. The U.S. has essentially no PR at the legislative level (a handful of cities have used variants for city council seats).

Each of these alternatives has different implications for third parties. RCV and approval voting reduce the spoiler problem at the candidate level: a voter can rank a Green candidate first and a Democrat second, knowing that if the Green is eliminated, the vote will transfer to the Democrat. PR eliminates the structural pressure entirely. We will return to all of these in §19.11.

19.5 The History of U.S. Parties: Six Party Systems (and a Disputed Seventh)

American party history is conventionally divided into six numbered "party systems," each demarcated by a major shift in the dominant cleavages, the geographic basis of party support, or the parties themselves. These divisions were articulated most influentially by political scientists V. O. Key, Walter Dean Burnham, and Paul Kleppner in the mid-twentieth century.⁴ Whether the divisions are as clean as the numbering suggests has been contested by recent scholarship, particularly Eric Schickler and Hans Noel.⁵ We will lay out the standard account, then note what has been challenged.

19.5.1 The First Party System (1796–1820s)

The first parties were the Federalists (associated with Alexander Hamilton, John Adams, and a strong-central-government, pro-merchant, pro-Britain orientation) and the Democratic-Republicans (associated with Thomas Jefferson and James Madison, with a strong-states, agrarian, pro-France orientation). The Federalists dominated through Adams's term (1797–1801); the Democratic-Republicans dominated through the Jefferson, Madison, and Monroe administrations (1801–1825). The Federalist Party effectively collapsed after 1815, partly because of its opposition to the War of 1812 and partly because the Hartford Convention of 1814 — at which some New England Federalists discussed possible secession — discredited the party nationally.

By the 1820s, the Democratic-Republicans were the only meaningful party, producing what is sometimes called the "Era of Good Feelings" but was in practice an era of intense factional competition within a single party.

19.5.2 The Second Party System (1828–1854)

The Second Party System emerged from the splintering of the Democratic-Republicans. Andrew Jackson's 1824 loss in the House (after winning a plurality of the popular and electoral vote, but failing to secure a majority — the "corrupt bargain" episode) led to the formation of the Democratic Party as Jackson's vehicle, and his 1828 election. Opponents of Jackson, who saw him as a populist demagogue with autocratic tendencies, organized as the Whigs. The Second Party System pitted Jacksonian Democrats (Westerners, Southerners, urban immigrants, advocates of expansion and limited federal economic intervention) against Whigs (Northerners, businesspeople, advocates of a national bank and internal improvements). The system collapsed in the early 1850s under the strain of the slavery question, which neither party could resolve internally.

19.5.3 The Third Party System (1854–1896)

The Whig Party fractured over slavery. Northern anti-slavery Whigs, joined by anti-slavery Democrats and former members of the short-lived Free Soil Party, founded the Republican Party in 1854. The new party was explicitly anti-slavery — at minimum opposing the extension of slavery to new territories, and, in many of its founders, committed to the eventual abolition of slavery itself.⁶

The 1860 election of Abraham Lincoln on the Republican ticket — with the South refusing to accept the result — triggered secession and the Civil War. After the war, Republicans dominated Reconstruction, the establishment of the Fourteenth and Fifteenth Amendments, and the federal protection of formerly enslaved people. The Compromise of 1877, ending Reconstruction in exchange for the disputed 1876 presidential election (Hayes-Tilden), began the long period of one-party Democratic rule in the South. From roughly 1877 through 1932, the South voted overwhelmingly Democratic, the North leaned Republican, and the parties oscillated in close national elections. This period is sometimes called the "Gilded Age" alignment.

19.5.4 The Fourth Party System (1896–1932)

The election of 1896 — William McKinley (Republican) defeating William Jennings Bryan (Democrat, also nominated by the Populist Party) — is the conventional marker of the start of the Fourth Party System. McKinley's victory consolidated Republican dominance outside the South, with industrial workers, farmers in the upper Midwest, and Northern cities aligning with the Republicans on a platform of high tariffs, the gold standard, and pro-business economic policy. The Democrats remained dominant in the "Solid South," but Republicans held the presidency for all but eight years (1913–1921, the Wilson administration) between 1896 and 1932.

19.5.5 The Fifth Party System (1932–1968)

The Great Depression and the 1932 election of Franklin Roosevelt produced the most consequential realignment in U.S. history. Roosevelt assembled the New Deal coalition: Northern urban workers, ethnic Catholics and Jews, African Americans (who shifted decisively from Republican to Democratic between 1932 and 1948), Southern whites (still Democratic for now), labor unions, and intellectual progressives. The Democratic Party became the dominant national party. Republicans held the White House for only eight years (Eisenhower, 1953–1961) of the thirty-six years between 1933 and 1969. The Democratic Party was internally heterogeneous — Northern liberals favored civil rights and an expanded welfare state; Southern Democrats opposed federal civil-rights enforcement and constituted a powerful conservative bloc within the party. The tension was contained for a generation but became unsustainable in the 1960s, after the Civil Rights Act (1964) and the Voting Rights Act (1965) — both passed with Republican support and against the opposition of much of the Southern Democratic delegation.

19.5.6 The Sixth Party System (1968–present? — disputed)

The 1968 election of Richard Nixon, with George Wallace's third-party candidacy capturing 13.5% of the popular vote and five Southern states, marked the beginning of a long realignment. Nixon's "Southern Strategy" — explicit in Nixon advisor Kevin Phillips's 1969 book The Emerging Republican Majority — sought to peel Southern white voters away from the Democratic Party, in part by appealing to opposition to federal civil-rights enforcement.⁷

Whether the Southern Strategy was the primary cause of Southern realignment, or whether other forces (the Voting Rights Act, the rise of suburban evangelicalism, economic restructuring of the South, the migration of Northerners to the Sun Belt) were more important, is contested by historians.⁸ What is uncontested is the result: by the late twentieth century, the South had become the most reliably Republican region in the country, the educated coastal cities had become the most reliably Democratic, and the parties had completed a nearly-total inversion of their nineteenth-century geographic bases.

The Reagan realignment of 1980 consolidated and extended this shift. Republicans assembled a coalition of fiscal conservatives, religious conservatives, Cold War hawks, and Southern and Western whites. Democrats remained dominant in the urban North and on the coasts but lost the South almost entirely.

The Sixth Party System has been the longest-running of the six. As of 2026, it is fifty-eight years old. Whether it has ended — and whether 2016 marked the start of a Seventh Party System — is one of the active debates in American political science.

19.5.7 The Disputed Seventh

The case for a Seventh Party System dates 2016 as a critical election: the year Donald Trump took over the Republican Party from outside the formal organization, the Democratic Party became more decisively the party of college-educated voters, the urban-rural divide deepened, and the parties realigned around education and density rather than primarily around the New Deal cleavages of class and labor.⁹

The case against treating 2016 as a critical election: the 1968-and-after coalitions did not actually collapse in 2016. The Republican Party kept its core Southern, evangelical, and rural base; it gained working-class non-college whites who had been drifting Republican since the 1990s but had still voted for Obama in 2008 and 2012. The Democratic Party kept Black voters, young voters, college-educated suburbanites, and a plurality of Hispanic and Asian voters; it lost some white working-class voters to Trump. These shifts are real but incremental. They look more like the gradual evolution of an existing party system than the kind of break-and-rebuild that 1860 or 1932 represented.¹⁰

The honest answer, as of 2026, is that we will probably need another decade of data before we can say with confidence whether 2016 was the start of a Seventh Party System or simply a particularly volatile late period of the Sixth. The numbering may matter less than the underlying coalitional shifts, which we describe in §19.6 and §19.7.

19.6 The Republican Party in 2026

The Republican Party of 2026 is, in important respects, a different party from the one that nominated Mitt Romney in 2012. It is also recognizably continuous with the party of Reagan and the post-1968 Sixth Party System. We will describe both the continuity and the transformation.

The 2024 election produced a Republican victory at the presidential level (Donald Trump's return to the White House), Republican Senate control (53 seats), and Republican House control (narrow). Trump's coalition was assembled from four overlapping groups:

Evangelical Protestants remain the most reliably Republican religious group. White evangelicals voted approximately 80% Republican in 2024, continuing a pattern that has held since the late 1990s. The Republican commitment to abortion restrictions, religious liberty (broadly defined), and traditional definitions of marriage and family make this alignment durable. The post-Dobbs (2022) landscape has, if anything, deepened it, even as some moderate suburban evangelicals have peeled off.

Working-class non-college whites — voters without four-year college degrees who identify as white — have shifted from a roughly even split in the 1990s and early 2000s to a 65–70% Republican preference by 2024. This shift was already underway during the Bush years, accelerated under Obama, and consolidated under Trump. Economic concerns (manufacturing employment, trade), cultural concerns (immigration, education-policy debates, what Republicans call "wokeness"), and geographic factors (rural and exurban residence) all contribute.

Younger Hispanic men, particularly in border states (Texas, Florida, Arizona), shifted toward Republicans in 2020 and 2024. Trump improved on his 2016 Hispanic vote share by nearly twenty points among male Hispanics under 45 in some swing-state subsamples. The reasons are debated: economic populism, religious conservatism (particularly among Pentecostal and evangelical Hispanics), reaction against progressive identity politics, immigration enforcement concerns among legal immigrants, and a generational shift in which younger Hispanics increasingly resemble the broader working class in their voting patterns.

Traditional pro-business and pro-military Republicans remain in the coalition but are no longer its center of gravity. The Reagan-era three-legged stool (fiscal conservatives, social conservatives, defense hawks) has been replaced by a coalition more focused on economic populism, immigration restriction, and cultural traditionalism. The Chamber of Commerce, once a reliable Republican ally on free-trade issues, broke with Trump on tariffs and immigration; many of its allied corporations have shifted donations toward Democrats. The neoconservative foreign-policy establishment — once a major Republican intellectual center — has largely been displaced by a more nationalist, less interventionist orientation.

The internal tensions within the Republican coalition are real and visible. They include:

National-conservative versus libertarian. The MAGA wing favors industrial policy, tariffs, restrictions on the free movement of capital across borders, and a more activist federal role in cultural questions (school curricula, pornography, immigration). The libertarian wing — represented by Senator Rand Paul, parts of the Cato Institute orbit, and the remaining Reagan-Goldwater conservatives — opposes both economic protectionism and federal cultural mandates. Both factions want lower taxes; they disagree on almost everything else.

Isolationist versus interventionist. The MAGA wing has been skeptical of NATO, of large-scale military aid to Ukraine, and of an American security commitment to Taiwan (though here positions vary). The traditional Republican foreign-policy establishment — Senators Mitch McConnell (until his retirement), Lindsey Graham, and the Reagan Foundation orbit — favors continued American leadership and alliance commitments. The 2024 vote on supplemental Ukraine aid divided the Republican caucus visibly.

Populist versus business-establishment. On taxation, the populist wing favors keeping or expanding the 2017 tax cuts for individuals while raising corporate rates; the business establishment favors the reverse. On regulation, both wings favor cuts, but in different sectors — the business establishment emphasizes financial deregulation and antitrust forbearance, the populist wing emphasizes tech-platform regulation and skepticism of large-corporate political spending. The Republican coalition's two halves have, in some respects, swapped traditional positions.

These tensions do not appear to be resolving. The 2028 Republican primary will be the first real test of whether the MAGA-aligned faction has become the new permanent center of the party (as the Reagan-aligned faction was the new permanent center after 1980) or whether the establishment-business wing can mount a comeback.

19.7 The Democratic Party in 2026

The Democratic Party of 2026 is the most demographically diverse coalition in the modern history of either party. It is also internally divided, in ways that the post-2024 election has made unusually visible.

The 2024 election produced a Democratic loss at the presidential level (Vice President Kamala Harris's defeat after President Biden's late-campaign withdrawal), continued Democratic minorities in both chambers of Congress (47 senators, 213 House members), and a depressed Democratic share among several groups that had been reliably Democratic for decades. The post-election autopsy debate is ongoing as of 2026.

The Democratic coalition includes:

College-educated professionals — white, Asian, Black, and Hispanic, with the white college-educated share showing the most dramatic Democratic shift since 2008. White college graduates voted 51% Democratic in 2024, up from 47% in 2008 and 42% in 1996. This is the single largest demographic shift in either coalition over the past three decades.

Most Black voters, particularly older Black voters and Black women. Black voters supported Democrats by approximately 80% in 2024, down modestly from approximately 85% in 2008, with most of the slippage among younger Black men.

Pluralities of Hispanic and Asian voters, with both groups showing modest movement toward Republicans (especially among male and younger voters). Hispanic voters supported Democrats 55–58% in 2024, down from 67% in 2012. Asian voters supported Democrats 60–63% in 2024, down from 73% in 2012.

Most LGBT-identifying voters, who have voted approximately 75–80% Democratic in every election since orientation has been measured.

Younger voters, with caveats. Voters under 30 supported Democrats by roughly ten points in 2024, but the youngest cohort (18–24) showed less Democratic advantage than the 25–29 cohort, suggesting the generational tide may have crested.

Union members, though this group is now smaller and less reliably Democratic than in the New Deal era. Public-sector union members remain reliably Democratic; private-sector union members (particularly building-trades unions) have shifted Republican enough that some endorsements went to Trump in 2024.

The internal tensions in the Democratic coalition are different from the Republican tensions but no less consequential:

Progressive versus moderate. The progressive wing — represented by Senator Bernie Sanders, the House "Squad" (Reps. Alexandria Ocasio-Cortez, Ilhan Omar, Rashida Tlaib, Ayanna Pressley, and others elected with their endorsement), and the Working Families Party — favors Medicare for All, a Green New Deal, expanded labor protections, and substantial wealth taxes. The moderate wing — represented by Senators Mark Kelly and Jon Tester (until his 2024 loss), the Problem Solvers Caucus, and the Center for American Progress — favors incremental expansion of the Affordable Care Act, decarbonization through tax incentives rather than mandates, and centrist-pragmatic messaging on cultural issues. The two factions agree on many specifics but differ in tone, ambition, and political theory.

Bernie left versus Clinton-Obama-Biden establishment. The Bernie Sanders 2016 and 2020 primary campaigns built a faction that has not gone away. It has elected representatives, captured state-level Democratic parties (in places), and produced a parallel media ecosystem (Jacobin, the Intercept, The Young Turks). The Clinton-Obama-Biden establishment — the institutional Democratic Party, the labor-aligned think tanks, the major Democratic donor network — has won most contested primaries since 2016 but has lost the cultural-energy battle in significant respects. The 2024 primary did not test this division because Biden faced no serious challenger. The 2028 primary will.

Identity-focused versus class-focused. The identity-focused wing argues that racial, gender, and sexual-orientation justice should be central to Democratic policy and messaging, and that the party's losses among working-class voters reflect the costs of being insufficiently committed to those goals, not too committed. The class-focused wing — represented by writers like Ruy Teixeira (who has since left the Democratic-aligned camp for the Center-Left "Liberal Patriot" project), David Shor's polling-driven analysis, and parts of the Joe Manchin / Bob Casey center-left — argues that the Democrats lost ground with working-class voters of all races by appearing to prioritize cultural issues over economic ones. The post-2024 autopsy has intensified this debate.

These tensions, like the Republican ones, do not appear to be resolving. The 2028 Democratic primary will, like the Republican one, be a major test of which faction commands the party's future.

19.8 Voter-Coalition Shifts Since 2008

The most important cross-cutting story of recent American politics is the shift in which kinds of voters support each party. We have referred to several pieces of this story already; this section pulls them together in one place.

The education realignment. The single largest demographic shift in voting behavior over the past two decades has been the divergence of college-educated and non-college voters. In 2008, the gap between four-year-college graduates and non-graduates in presidential vote share was approximately five points (with graduates leaning slightly more Democratic). By 2024, the gap was nearly twenty points among white voters and double-digit even within most racial groups. College-educated voters of every race now lean more Democratic than their non-college peers; non-college voters of every race now lean more Republican than their college peers. The shift has been particularly dramatic among white voters: white college graduates moved from R+11 in 2000 to D+5 in 2024, a sixteen-point swing. White non-college voters moved from R+18 to R+30 in the same period, a twelve-point swing in the opposite direction.

The racial realignment. Less dramatic than the education realignment, but real, has been a modest movement of Hispanic and Asian voters toward Republicans, particularly among men and younger voters. Hispanic vote share for Republicans rose from approximately 28% in 2012 (Romney) to approximately 42% in 2024 (Trump). Asian-American Republican share rose from approximately 27% to 36% in the same period. The reasons are debated: economic populism, religious conservatism, immigration-enforcement preferences among legal immigrants, generational generational assimilation patterns, and reaction to progressive identity politics among non-white voters who do not identify with that politics. Black voters showed a smaller shift, with younger Black men moving Republican more than older Black voters or Black women.

The gender gap. The persistent Democratic advantage among women, and Republican advantage among men, has held for decades but widened in 2024. In 2024, women voted Democratic by approximately 55–43; men voted Republican by approximately 55–43. The gap is roughly twelve to fourteen points. It is consistent across most racial groups.

The geographic realignment. The urban-rural divide has deepened. Counties with urban density above 1,000 residents per square mile voted Democratic by 30+ points in 2024; counties with density below 50 voted Republican by 40+ points. The suburban swing — once a major contested zone — has split, with college-educated suburbs trending Democratic and exurban areas trending Republican. The realignment has been clearest in the Northeast and Midwest, where the old industrial-suburban Democratic strongholds have become less Democratic and the old Republican upper-Midwest farms have become more Republican.

These shifts are not "the parties trading places" in any simple sense. The Republican Party of 2024 is not the Democratic Party of 1968. But the underlying coalitional bases have shifted enough that political-science textbooks written before 2008 will read, today, as describing a different country.

19.9 Party Organization

We turn now to the party-as-organization. The formal structures of the Republican and Democratic Parties are surprisingly similar, and surprisingly weak compared to the parties of mid-twentieth-century America.

The national committees. The Republican National Committee (RNC) and the Democratic National Committee (DNC) are the formal national-party organizations. Each has a chair, members representing each state and territory, a budget in the high hundreds of millions of dollars per cycle, and headquarters in Washington. Their formal powers include adopting party platforms (at the quadrennial conventions), recognizing state-party affiliations, governing the rules for delegate selection in presidential primaries, and disbursing party funds.

In practice, the actual influence of the national committees has declined since the mid-twentieth century. In the era of party machines, national committees were powerful brokering institutions. In the candidate-centered era — beginning roughly with the 1972 reforms (described in §19.10) and accelerating through the post-2010 small-donor-internet era — the national committees have become primarily fundraising and coordination operations rather than candidate-selection institutions. The RNC could not stop Donald Trump from being nominated in 2016, even though many of its members opposed him. The DNC could not stop Bernie Sanders from competing seriously in 2016 and 2020 (though the post-2016 reforms reduced superdelegate influence in ways that arguably hurt the establishment in 2020). The national committees fundraise, coordinate messaging, and run convention logistics. They do not, in general, decide who the nominee will be.

The campaign committees. The four congressional campaign committees — the National Republican Congressional Committee (NRCC, House Republicans), the Democratic Congressional Campaign Committee (DCCC, House Democrats), the National Republican Senatorial Committee (NRSC, Senate Republicans), and the Democratic Senatorial Campaign Committee (DSCC, Senate Democrats) — are far more important to the day-to-day political operations of the parties than the national committees are. Each congressional campaign committee is staffed by paid professionals, raises tens to hundreds of millions of dollars per cycle, recruits candidates, runs advertising, polls competitive races, and provides intelligence to incumbents and challengers. These committees, more than the RNC and DNC, are where the party-as-organization actually does political work.

State parties. The fifty Republican state parties and fifty Democratic state parties (plus the DC and territorial affiliates) vary enormously in capacity. Some — the Wisconsin Democrats, the Texas Republicans, the California Democrats — are large, professional, well-funded organizations that can rival small federal agencies in size. Others — the Wyoming Democrats, the Massachusetts Republicans — are skeleton operations with handfuls of staff. State parties handle ballot-access, primary administration (in coordination with state election authorities), voter contact, and the kind of long-term organization-building that national committees cannot do at the local level.

Local parties. Local party organizations — county committees, ward committees, precinct captains — were once the engine of American politics. The Cook County Democratic Organization (the "Daley machine," for those familiar with twentieth-century Chicago) provided patronage, voter contact, and turnout operations for generations. By the 2020s, this kind of organization is mostly gone. Local parties exist on paper in most counties; their actual day-to-day political activity is, in most places, modest. The death of the local party is a significant feature of the modern American political landscape and a recurring concern of analysts who worry about civic disengagement.¹¹

The result is a party system in which the formal organizations are real and consequential but considerably less powerful than the underlying party labels and the candidate-centered campaigns that operate beneath the labels.

19.10 The Primary System

The most consequential change to American party politics in the past century was the transformation of the presidential nomination process — specifically, the post-1972 shift from convention-centered nomination by party leaders to primary-centered nomination by primary voters.

19.10.1 Pre-1972

Before 1972, primaries existed but did not determine nominees. Most delegates to national conventions were chosen by party leaders — state-party chairs, governors, members of Congress, "bosses" of urban political machines — through caucuses, state conventions, and informal negotiation. Primaries existed in some states (New Hampshire's primary was famously consequential) but functioned as performance opportunities for candidates to demonstrate viability rather than as binding commitments. The 1968 Democratic Convention, held in Chicago amid riots and protests, became a turning point. Hubert Humphrey, the establishment candidate, secured the nomination without having competed in primaries; activists who had backed Eugene McCarthy and Robert Kennedy (assassinated in June 1968) regarded the process as illegitimate. The fallout produced the McGovern-Fraser Commission, which rewrote the Democratic Party's delegate-selection rules to require open, primary-driven selection.

19.10.2 Post-1972

The McGovern-Fraser reforms, adopted by Democrats in 1972 and substantially copied by Republicans in subsequent cycles, created the modern primary system. State parties (in coordination with state election authorities) hold primaries or caucuses. The candidates who win those primaries accumulate delegates. The convention ratifies what the primary voters have decided. Party leaders' formal role in selecting nominees is reduced to that of "superdelegates" (in the Democratic Party — the Republicans have no superdelegates) or "uncommitted" delegates whose role is largely ceremonial.

The consequences of this transformation have been debated for fifty years. Defenders of the post-1972 system argue that it democratized party nominations, opening the process to broader participation and reducing the power of unaccountable party bosses. Critics argue that it removed the party-as-organization's gatekeeping function, allowing candidates to win nominations without the support of (and sometimes against the preferences of) the party-as-organization. The most influential statement of the critic's position is Marty Cohen, David Karol, Hans Noel, and John Zaller's The Party Decides: Presidential Nominations Before and After Reform (2008), which argued that party insiders — donors, elected officials, activists, and key constituency leaders — still effectively coordinated to select nominees in the post-1972 era, even without formal control of the convention.¹²

The "party decides" thesis was influential in the 2010s but was tested severely in 2016, particularly by the Republican primary. We turn to that test in case-study-01.

19.10.3 The Invisible Primary

Even in the post-1972 system, much of the actual selection happens before any primary votes are cast. The "invisible primary" is the year-long competition among candidates to secure donor commitments, key endorsements (governors, senators, big-city mayors), positive media coverage, and early-state organizational presence. Candidates who fail to compete in the invisible primary are almost always eliminated before primary voting begins. Hillary Clinton's 2016 invisible-primary dominance gave her a structural advantage over Bernie Sanders that survived the entire primary season. Joe Biden's 2020 invisible-primary performance was, by contrast, mediocre — he was widely written off after Iowa and New Hampshire — but his support among Black voters in South Carolina, combined with a coordinated establishment endorsement after his South Carolina victory, demonstrated that the invisible primary was still operating, just with different actors than in previous cycles.

19.10.4 Open, Closed, and Semi-Closed

State primary rules vary along a spectrum. Closed primaries allow only registered party members to vote in that party's primary; open primaries allow any registered voter to participate in any party's primary, regardless of registration; semi-closed primaries allow party members and unaffiliated voters to participate but exclude registered members of other parties. The mix matters because it changes the electorate: closed primaries tend to produce more ideologically committed nominees; open primaries can be influenced by independents and even, occasionally, by mischievous cross-party voters. As of 2026, approximately fifteen states have closed primaries, twenty have open primaries, and the rest fall somewhere in the semi-closed range. Each state party can, within limits, choose its own primary type for federal races, and the rules sometimes change between cycles.

19.10.5 The 2016 and 2020 Tests

The 2016 Republican primary, examined in detail in case-study-01, was the most severe test the "party decides" thesis has faced. The 2020 Democratic primary — Joe Biden's establishment-backed coalescence after a poor early-state showing — was, by contrast, evidence that the thesis still held, at least under certain conditions. We will not duplicate that case study here. The point for the structure of this section is that the post-1972 primary system has not eliminated the party's role in nominations, but it has made that role contestable in ways that the pre-1972 system did not allow.

19.11 Why Third Parties Don't Win

Throughout this chapter we have referred to the persistent failure of third parties in U.S. politics. We now examine that failure systematically. The structural reasons are familiar from §19.3 (Duverger plus institutional barriers). The historical record is worth reviewing because it tells us something about how third-party movements end — and about the conditions under which third-party energy actually produces change.

The major third-party presidential candidacies of the post-WWII era:

George Wallace, 1968 (American Independent Party): 13.5% of the popular vote; 46 electoral votes (carried five Southern states). The most successful third-party performance since 1968. Wallace's coalition — Southern segregationists, working-class whites, voters opposed to federal civil-rights enforcement — was largely absorbed by the Republican Party in the 1970s and 1980s.

John Anderson, 1980 (independent): 6.6% of the popular vote, zero electoral votes. Anderson, a moderate Republican, ran as an independent after losing the Republican primary to Reagan. His support faded by Election Day; the spoiler effect was minimal because Reagan won decisively.

Ross Perot, 1992 (independent, then 1996 Reform Party): 18.9% of the popular vote in 1992 — the largest popular-vote share by a third-party candidate since Theodore Roosevelt in 1912 — and 8.4% in 1996. Perot did not win any electoral votes in either election. His 1992 candidacy is widely credited with contributing to George H. W. Bush's loss to Bill Clinton, though political scientists who have studied the question carefully argue Perot drew roughly equally from Bush and Clinton voters, so the spoiler effect is contested. Perot's issues — fiscal responsibility, opposition to NAFTA, "the giant sucking sound" of jobs leaving — were partially absorbed by both parties in the 1990s.

Ralph Nader, 2000 (Green Party): 2.7% of the popular vote, zero electoral votes. Nader's candidacy is a paradigm case of the spoiler effect in U.S. politics. In Florida, Nader received 97,488 votes; George W. Bush defeated Al Gore in Florida by 537 votes. The spoiler argument: had Nader not been on the ballot, enough of his voters would have supported Gore to swing Florida and the election. The defense: Nader argued (a) some Nader voters would not have voted at all without him on the ballot, (b) some would have voted for Bush, and (c) Gore's loss was caused by other factors (the butterfly ballot, the media calling Florida for Gore prematurely, Gore's own campaign decisions). The empirical evidence suggests Nader did, on net, cost Gore Florida and the presidency, but the magnitude was small relative to the multiple other contingencies of that election.¹³

Gary Johnson, 2016 (Libertarian): 3.3% of the popular vote, zero electoral votes. Jill Stein, 2016 (Green): 1.1% of the popular vote, zero electoral votes. Stein's vote totals in Michigan, Pennsylvania, and Wisconsin exceeded Trump's margins of victory in those states, leading to spoiler arguments from Democrats. The empirical analysis is complicated by the fact that some Stein voters would not have voted at all in a two-way race.

Robert F. Kennedy Jr., 2024 (independent, then suspended and endorsed Trump): Polling above 10% nationally in early 2024; finished at approximately 0.6% after a late suspension. His campaign trajectory illustrates how third-party candidacies typically erode as Election Day approaches and voters return to the major parties.

The pattern across all of these is consistent. Third-party candidates can affect outcomes (by drawing votes from one major-party candidate more than the other) but cannot win. They can put issues on the agenda that the major parties have been ignoring (Perot on the deficit, Nader on consumer protection and corporate power) and sometimes succeed in getting those issues absorbed by a major party. They occasionally produce major-party realignments by exposing coalitional weaknesses (Wallace's 1968 candidacy revealed the depth of Democratic vulnerability in the South). What they do not do, in the contemporary U.S. system, is win the presidency, win Senate seats, or win House seats in any sustained way.

The success of factional movements within major parties stands in instructive contrast. The Tea Party movement (2009–2014) did not become a third party; it ran candidates in Republican primaries, won many of them, and reshaped the Republican Party from the inside. The democratic-socialist movement around Bernie Sanders did not become a third party; it has elected representatives in Democratic primaries (the Squad, plus state-level Democratic socialists in Michigan, New York, and elsewhere) and pushed the party's policy agenda leftward. The MAGA movement did not become a third party; it captured the Republican Party from inside in 2016 and has held it since.

The pattern suggests something important about American political reform energy. In a multi-party system, factional movements form their own parties; in the U.S. system, they fight for control of one of the two existing parties. Whether this is good or bad — whether absorbing factional energy into existing institutions is healthier than letting it form its own — is a question the reform proposals in §19.11 try to address.

19.12 Reform Proposals

If the U.S. two-party system is producing outcomes that critics find unsatisfactory — high polarization, candidates whom most voters do not love, persistent third-party failure, and primary nominations that produce general-election candidates who appeal to narrow primary electorates — what could be done to change it?

We will examine four classes of reform: Ranked-Choice Voting, approval voting, top-two primaries, and proportional representation. We will steel-man each, including the case for the status quo.

19.12.1 Ranked-Choice Voting (RCV)

The case for: RCV reduces the spoiler problem by allowing voters to rank candidates rather than choose only one. A voter who prefers a Green candidate but would settle for a Democrat over a Republican can rank Green first and Democrat second; if the Green is eliminated, the vote transfers to the Democrat. This eliminates the strategic-voting pressure that suppresses third-party support under FPTP. RCV also tends to produce winners with broader majority support, since the eventual winner is the candidate who can build a coalition across multiple preference rankings rather than the one who wins a plurality of first-place votes. Maine's 2018 election (the Bruce Poliquin / Jared Golden race in ME-2) and Alaska's 2022 special election (Mary Peltola's victory over Sarah Palin) are recent examples.

The case against: RCV is more complex than FPTP. Voters in jurisdictions that have adopted RCV report higher rates of confusion, particularly among older voters and voters with limited formal education. RCV can produce "non-monotonic" outcomes in some rare cases (where voting for a candidate higher can cause that candidate to lose, contrary to intuition). The Burlington, Vermont 2009 mayoral election produced what RCV critics call a paradoxical result: a candidate who could have beaten any other candidate head-to-head (the "Condorcet winner") lost, while a different candidate won under the RCV elimination rules. Burlington repealed RCV the next year (the city later restored it for some elections). Alaska's 2022 special election was contested by Sarah Palin's campaign on similar grounds; supporters argued the result reflected the broader preferences of Alaska voters across all rankings, which is exactly what RCV is supposed to capture.

The case for the status quo: FPTP is simple, transparent, and immediately understandable to every voter. RCV adds complexity that may, in close cases, produce results that voters cannot easily explain or trust. In a low-trust environment, this matters.

19.12.2 Approval Voting

The case for: Approval voting asks voters to vote for as many candidates as they approve of, then awards the seat to the candidate with the most approvals. It is simpler than RCV (no ranking required) and reduces the spoiler problem by allowing voters to support multiple candidates simultaneously. Fargo, North Dakota (2018) and St. Louis, Missouri (2020) have adopted approval voting; early empirical studies suggest voter satisfaction with the system is high.

The case against: Approval voting has a much shorter empirical track record than RCV in U.S. elections. The strategic-incentive analysis is also more complex than it appears: voters have to decide where to set their "approval threshold," and that decision interacts with their predictions about other voters' choices in complicated ways. Some game-theoretic analyses suggest approval voting can collapse to FPTP when voters anticipate close two-candidate races.

19.12.3 Top-Two and Top-Four Primaries

The case for: Open primaries that put all candidates from all parties on a single ballot, advancing the top finishers regardless of party, reduce the influence of partisan primary electorates and create incentives for general-election candidates to appeal to the median voter. California's top-two system has been credited (by some analyses) with reducing the number of extreme nominees produced by the parties. Alaska's 2020 reform combines a top-four primary with an RCV general election, producing what reformers argue is the best-of-both-worlds combination.

The case against: Top-two primaries can produce same-party general elections in safe districts, depriving voters of meaningful choice between parties. In a heavily Democratic California district, both general-election candidates may be Democrats; in a heavily Republican district, both may be Republicans. Critics argue this is a feature (the choice is now on the dimensions on which Democrats or Republicans differ from each other within the party); defenders of the status quo argue it is a bug (general-election voters lose the option to vote for the other party).

19.12.4 Proportional Representation (PR)

The case for: PR with multi-member districts directly addresses the structural pressure of Duverger's Law. A party with 15% of the vote gets 15% of the seats. Third parties become viable; coalitions become explicit and negotiated rather than formed before the election by the party-as-organization. Most consolidated democracies use some form of PR; their politics is generally less polarized than American politics (though the causal direction is debated).

The case against: PR requires multi-member districts, which would be a constitutional and legal earthquake for Congress (state legislatures could move first). PR tends to produce coalition governments, which can be unstable (Italy, Israel) or can create kingmaker positions for small extreme parties that disproportionately influence policy. PR breaks the link between a single representative and a single geographic constituency, which many Americans value. The case for PR is, in important respects, a case for changing the basic structure of American representation in ways that go far beyond party reform.

19.12.5 The Case for the Status Quo

It is worth giving the strongest version of the conservative-of-the-system case as well. The American two-party system has produced:

• Over 230 years of continuous democratic governance under a single constitutional framework, longer than any other large democracy.
• Two clear teams that voters can hold accountable through alternation in office.
• A structural pressure toward broad-tent coalitions that, in theory, force each party to negotiate internally rather than externally.
• Stability in legislative governing majorities (compared to the coalition reshuffling of multi-party PR systems).
• Predictability for voters about what each party will do if elected (compared to multi-party coalition negotiations whose outcomes voters cannot predict in advance).

Critics of the current system argue that these advantages have eroded — that polarization, gridlock, and institutional dysfunction have undermined whatever benefits two-party systems were once supposed to deliver. Defenders argue that the American system has weathered worse and that the proposed cures may be worse than the diseases.

We will not, in this textbook, take a position on which assessment is correct. The honest answer is that reasonable people disagree about whether the U.S. party system needs structural reform or merely needs to function within its existing structure. The empirical and theoretical evidence on each reform is genuinely mixed. What is clear is that the structural pressure of Duverger's Law will continue to produce two major parties as long as single-member-district plurality voting remains the dominant electoral system. Anyone who wants more than two parties in the United States needs to change that voting system. Anyone who is satisfied with two parties — even while wanting different two parties, or different policies within them — does not.

19.13 What This Chapter Has Done

This chapter introduced political parties as institutions, as labels, and as governing coalitions. It explained the structural pressure of Duverger's Law and the additional institutional barriers that reinforce two-party dominance in the United States. It traced six (or seven) party systems in U.S. history and noted the active scholarly debate over whether 2016 marked a critical realignment. It described the early-2026 Republican and Democratic coalitions in detail, including the genuine and ongoing internal tensions within each. It catalogued the major demographic shifts since 2008 — the education realignment, the modest racial realignment, the persistent gender gap. It examined the post-1972 primary system and the contested "party decides" thesis. It explained why third parties consistently fail to win major office while factional movements within parties succeed. And it steel-manned each of the major reform proposals — Ranked-Choice Voting, approval voting, top-two primaries, proportional representation — alongside the case for the status quo.

The next chapter (Chapter 20) examines elections and campaigns: the rules by which we choose officeholders, the strategies campaigns deploy, and the role of money. Chapter 21 follows with the operational details of how a modern campaign actually runs. The party institutions described in this chapter are the platforms on which all of those campaign activities sit.

We close with a reminder that has run through every chapter of this book: the system was designed for disagreement. The two-party structure is one specific institutional form that disagreement has taken in the United States. It has both costs and benefits, and reasonable Americans across the ideological spectrum disagree about which dominate. The job of citizens — the job we will return to in Part V — is not to demand that the system produce outcomes any particular faction prefers. It is to understand the system clearly enough to operate inside it effectively, and to argue for changes to it from a position of understanding rather than grievance.