Chapter 35 — Gerrymandering and Electoral Design: When Politicians Choose Their Voters

"The voters don't choose the politicians. The politicians choose the voters." — common phrasing of the gerrymandering critique, attributed in various forms to reformers since at least the 1980s.

"There are no neutral lines." — Bernard Grofman, political geographer, who has spent forty years measuring redistricting and reminds his students that somebody draws every map and every drawing rule favors someone.

35.1 The shape of the problem

In 1812, Massachusetts governor Elbridge Gerry signed a redistricting bill drawn to favor his Democratic-Republican Party. One of the new state-senate districts north of Boston resembled, in the cartoonist's eye, a salamander. The Boston Gazette coined the portmanteau "Gerry-mander." The word entered the language. The practice was older than the word and is, two centuries later, the central institutional pathology of American legislative elections.

This chapter is about that pathology — and about the things people sometimes confuse with it. Three claims structure what follows, and each requires defense.

Claim one: Gerrymandering is real, measurable, and consequential. A drawn map can shift the partisan composition of a state's congressional delegation by five to ten seats. Whichever party controls the pen draws the map; whichever party does not gets the table scraps. The mechanics — packing and cracking — are not subtle. The empirical effect on competitive elections is not subtle. Citizens who care about the responsiveness of legislatures to voters should care about how lines are drawn.

Claim two: Gerrymandering is not the only — or even, by some measures, the largest — driver of safe seats in Congress. Geographic sorting matters too. Republican voters cluster in rural counties where Republican vote shares run 70–80%; Democratic voters cluster in urban counties where Democratic vote shares run 75–90%. Even neutral, computer-drawn maps in a single-member-district plurality system tend to produce many lopsided districts because the underlying voters are lopsidedly distributed in space. Reformers who promise that better maps will deliver wholesale competitive elections oversell what redistricting reform can do.

Claim three: Both major parties gerrymander when they control the pen, and they have done so for as long as the modern redistricting cycle has existed. This is empirically true. It is also, in 2026, true with asymmetries that the data should report frankly: in the 2010 cycle, Republicans drew more aggressive maps in more states than Democrats did, producing a measurable seat advantage in the U.S. House for most of the decade. In the 2020 cycle, the asymmetry narrowed, partly because Democratic-controlled states (New York, Maryland, Illinois) drew aggressive maps of their own, partly because state courts struck down both Republican (North Carolina, before reversal) and Democratic (New York) maps, and partly because the Voting Rights Act forced redrawing in Alabama. This chapter will name the maps, the mapmakers, and the years.

You should leave this chapter able to (a) define gerrymandering and distinguish its types, (b) describe how districts get drawn in the United States and which states use which method, (c) apply the legal framework — the cases, the doctrines, the remaining causes of action after Rucho — to a specific map, (d) interpret quantitative measures of partisan bias such as the efficiency gap and compactness scores, (e) evaluate the most-prominent reform proposals against their strongest critiques, and (f) audit the history of your own congressional district as part of the Democracy Audit running project.

35.2 Definitions: the terms of art

Before any analysis, the vocabulary needs to be precise.

Redistricting is the lawful, scheduled redrawing of legislative district lines after each decennial Census, required by the principle of one person, one vote (Section 35.4). It is unavoidable and not, in itself, gerrymandering.

Gerrymandering is the strategic drawing of lines to advantage a party, an incumbent, an ethnic group, or a faction. Gerrymandering is a kind of redistricting; it is not a synonym.

Partisan gerrymandering draws lines to favor one major party. The line-drawer typically controls a state legislature and uses partisan voter-registration data, prior-election results, and demographic data to optimize seats won under expected partisan swings.

Racial gerrymandering draws lines based on race. The phrase has two distinct meanings that the law treats differently, and confusing them invites legal and analytical mistakes:

• Race-based dilution — drawing lines that dilute a racial minority's voting power by packing them into one district or cracking them across several. This is prohibited by Section 2 of the Voting Rights Act of 1965, when certain conditions are met (Section 35.4 below).
• Race-as-the-predominant-factor maps — drawing lines where race is the predominant line-drawing principle, even if the goal is to increase minority representation. This is subject to strict scrutiny under the Equal Protection Clause of the Fourteenth Amendment (Shaw v. Reno, 1993).

The interaction between these two doctrines is genuinely difficult. The Voting Rights Act sometimes requires mapmakers to consider race; the Constitution sometimes forbids them from making race predominant. Mapmakers have to navigate the gap. The Supreme Court's more recent decisions (especially Allen v. Milligan, 2023) clarified that there is, in fact, a workable path through the gap, contrary to predictions of some scholars in the 2010s.

Incumbent-protection gerrymandering draws lines to protect sitting members regardless of party. It is often bipartisan: the parties' legislative leaders agree to a map that makes most incumbents safe at the cost of competition. Congressional district maps in California before 2010, and in many state legislatures historically, have been bipartisan incumbent-protection maps.

The two basic moves of partisan gerrymandering are easy to picture:

• Packing. Pack as many of the opposing party's voters as you can into a small number of districts where they win by 80–90%. Their votes are "wasted" — they win, but only one seat.
• Cracking. Crack the rest of the opposing party's voters across several districts where your own party has a comfortable but not dominant majority — say, 55–60%. Your candidates win each of those seats.

A skillful gerrymander combines packing and cracking so that the opposing party "wastes" votes on both sides of the equation: they pile up unnecessary votes in their packed seats and they fall short by a few percentage points in many cracked seats. The efficiency gap (Stephanopoulos and McGhee 2014) is one quantitative measurement of this combined wasting; we examine it in Section 35.5.

A clean illustration. Imagine a state with 1,000,000 voters split exactly 50-50 between Party A and Party B, and ten congressional districts of 100,000 each. With perfectly neutral lines you might expect five seats each. But the line-drawer for Party A could, in principle, design four districts where Party B wins 80–20 (200,000 of A's opponents packed into one-fifth of the seats) and six districts where Party A wins 55–45 (the rest of A's opponents cracked across the remaining six). Party A wins 6 of 10 seats with 50% of the vote. Same voters, same total ballots cast, different lines. That is the basic geometry.

Real maps rarely look that clean — the underlying geography of voters constrains the line-drawer, the courts impose contiguity and equal-population requirements, and the Voting Rights Act may require certain configurations. But the basic logic operates in real maps. The 2010-cycle North Carolina, Pennsylvania, Wisconsin, and Ohio maps produced exactly this kind of disconnect between statewide vote totals and seat shares. Subsequent court rulings, redrawings, and a 2018 Pennsylvania state-supreme-court decision reduced (though did not eliminate) the asymmetry.

35.3 The redistricting process: who actually draws the lines

The Constitution, Article I, Section 4, leaves the Times, Places and Manner of holding congressional elections to state legislatures, subject to federal override. Most states follow the default: the state legislature draws congressional and state-legislative lines, the governor signs or vetoes (in most states), and the courts adjudicate challenges. But a growing number of states have moved to commissions or other arrangements, and the variation is now wide enough that students should understand the menu.

Legislature draws (default). Most states. The majority party's legislative leadership writes the map, with input from caucus staff, outside redistricting consultants (firms like the conservative-aligned Public Opinion Strategies or the progressive-aligned Strategic Telemetry), and the state demographer. A floor vote ratifies the map; the governor signs or vetoes. In states where one party controls the legislature and the governorship, the governor's veto is rarely a real check; in states with split control, divided maps are more likely to result. Examples in the 2020 cycle: Texas (Republican legislature and Republican governor, aggressive Republican map), Illinois (Democratic legislature and Democratic governor, aggressive Democratic map), Florida (Republican legislature and Republican governor; the governor pushed for a more aggressive map than the legislature initially proposed and got it), Maryland (Democratic legislature drew an aggressive map; struck down by a state court).

Bipartisan commission. A commission with members appointed by both parties' leadership, designed to require cross-party agreement. Examples: New Jersey (state-legislative; congressional commission has tie-breaking member); Washington State; Connecticut.

Independent commission. A commission whose members are insulated from elected officials by design — selected through application, screening by retired judges, lottery, or some combination, with rules that limit recent partisan activity. Examples:

• Arizona, after a 2000 ballot initiative — the first state to fully empower an independent commission for both congressional and state-legislative maps.
• California, after Propositions 11 (2008, state-legislative) and 20 (2010, congressional). The California commission has fourteen members — five Democrats, five Republicans, four unaffiliated — and requires a supermajority including votes from each subgroup to approve a map.
• Michigan, after a 2018 ballot initiative. Thirteen members chosen by lottery from applicants screened for partisan-affiliation balance.
• Colorado, after 2018 ballot measures (Amendments Y and Z).
• Virginia, after a 2020 constitutional amendment created a hybrid commission with legislator and citizen members.

Non-partisan staff drafting (with legislative ratification). Iowa is the textbook case. The state's Legislative Services Agency drafts a map according to neutral criteria (compactness, county integrity, no consideration of partisan or incumbent data); the legislature votes the map up or down without amendment for up to three rounds. After three rejections, the legislature can amend. In practice, Iowa rarely needs round three. The Iowa system has produced congressional maps with high compactness and no partisan-bias detectable by the standard measures since the early 1980s. Iowa's geography (smaller, denser population, fewer competing constituencies) makes the system easier to operate than it would be in a large or geographically complex state.

Other arrangements. Some states use a hybrid: legislature drafts but a commission has backstop authority; or commission drafts and legislature ratifies; or various tie-breaking rules. Maine, Idaho, and others have backup commissions that activate if the legislature fails to act on time.

The governor's veto. In most states the governor can veto a redistricting bill the same way they would any other bill. In a handful of states (notably North Carolina), the governor has no veto over redistricting. This was significant in North Carolina in 2011 and again in 2021–22, when Democratic governors lacked any formal check on Republican-legislature maps.

Census timing and the 2020 cycle. Redistricting follows the decennial Census. The 2020 Census, conducted partly under pandemic conditions and amid disputes about citizenship questions and undercount adjustments, was delivered to the states in 2021 — later than usual, compressing the redistricting timeline. Reapportionment of the 435 House seats among the states reflected the population shifts of the 2010s. Texas gained two seats; Florida, North Carolina, Colorado, Montana, and Oregon each gained one. New York, Illinois, Michigan, Ohio, Pennsylvania, West Virginia, and California each lost one. Several of those changes had partisan implications (red states gaining, blue states losing) and several had cross-cutting effects (Colorado and Oregon, both gaining, are blue states). The full apportionment table is on the Census Bureau's website; the per-state seat changes shape much of the 2022 and 2024 election analysis.

When you read about who controls redistricting in your state, the most-relevant facts are: which body draws (legislature, bipartisan commission, independent commission, staff), what is the governor's role, what are the constitutional and statutory criteria for the map, what state-court doctrine has emerged on partisan or racial gerrymandering. Section 35.7's Democracy Audit asks you to look these up.

35.4 The legal framework: constitutional and statutory limits

Five doctrines, six cases, and one statute frame what mapmakers can and cannot do. Each requires a brief tour.

One person, one vote. Two 1964 cases established that legislative districts must be roughly equal in population. Reynolds v. Sims (1964) applied the rule to state legislative districts; Wesberry v. Sanders (1964) applied it to U.S. congressional districts. The doctrine derives from the Equal Protection Clause and from Article I, Section 2 (which says representatives are apportioned "according to their respective Numbers"). For congressional districts the standard is essentially zero population deviation; for state legislative districts the standard allows up to 10% deviation between largest and smallest district. The doctrine ended an era in which rural districts had vastly disproportionate representation in many state legislatures. It is the bedrock; nothing else in the redistricting law makes sense without it.

Voting Rights Act, Section 2. Congress passed the Voting Rights Act of 1965 as a centerpiece of the civil-rights statutory revolution. Section 2, as amended in 1982 and interpreted in Thornburg v. Gingles (1986), prohibits voting practices and procedures (including redistricting plans) that result in discrimination on the basis of race, color, or membership in a language-minority group. To make a Section 2 vote-dilution claim, plaintiffs must establish three preconditions (the Gingles factors): the minority group is sufficiently large and geographically compact to constitute a majority in a single-member district; the minority group is politically cohesive; and the white majority votes sufficiently as a bloc to usually defeat the minority group's preferred candidate. If those preconditions are met, the court evaluates the totality of circumstances.

Voting Rights Act, Section 5 (preclearance). From 1965 until 2013, Section 5 required certain jurisdictions (mostly Southern states with histories of racial voting discrimination) to preclear any voting-related changes — including redistricting plans — with the U.S. Department of Justice or a federal court before implementing them. Shelby County v. Holder (2013) struck down Section 4(b)'s coverage formula, effectively ending preclearance. After Shelby County, Section 2 became the principal Voting Rights Act tool for challenging maps; the burden shifted from the jurisdiction (under preclearance, the jurisdiction had to prove non-discrimination) to the plaintiff (under Section 2, the plaintiff has to prove discrimination). The change is consequential.

Strict scrutiny under the Equal Protection Clause. Shaw v. Reno (1993) held that maps in which race is the predominant factor in line-drawing are subject to strict scrutiny under the Fourteenth Amendment. The case involved a North Carolina district drawn to be majority-Black that snaked along Interstate 85 for hundreds of miles. The Court held that such a map, where race is the predominant factor, must be narrowly tailored to a compelling state interest. Compliance with the Voting Rights Act is one compelling state interest, but only if the VRA actually requires the configuration drawn. In the years after Shaw a series of cases — Miller v. Johnson (1995), Bush v. Vera (1996), Cooper v. Harris (2017) — fleshed out the doctrine.

Vieth v. Jubelirer (2004) — partisan gerrymandering claims unmanageable. The Supreme Court split 4-1-4 on the question whether partisan-gerrymandering claims were justiciable — that is, whether courts could decide them at all. Four justices said no. Four said yes. Justice Anthony Kennedy, in the controlling concurrence, said yes in principle but no in practice until somebody developed a "judicially manageable standard." For fifteen years, lawyers and political scientists tried to build that standard.

Rucho v. Common Cause (2019). Chief Justice John Roberts, writing for a 5-4 majority, held that partisan-gerrymandering claims are non-justiciable political questions in federal court. The decision did not say that gerrymandering is constitutional. It said that federal courts should not adjudicate it because no manageable standard exists. Roberts pointed to other potential remedies: state courts under state constitutions, state legislation, ballot initiatives establishing commissions, federal legislation. He left those avenues open. The decision was a profound disappointment to the reformers who had spent the prior decade developing measures (efficiency gap, partisan symmetry, mean–median difference); it was a relief to mapmakers and party leaders who feared a flood of litigation. Either way, Rucho is the federal-court bottom line: federal judges will not strike down maps for being too partisan.

State courts and state constitutions. Rucho explicitly invited state-court action under state constitutions, and several states accepted the invitation. Pennsylvania: League of Women Voters v. Commonwealth (Pa. 2018) held that the 2011 Republican-drawn congressional map violated the Free and Equal Elections Clause of the Pennsylvania Constitution, struck the map down, and substituted a court-drawn map. The decision became a model. North Carolina: Harper v. Hall (N.C. 2022) struck down Republican-drawn maps under the state constitution; after the November 2022 election flipped the state supreme court's partisan composition, the court reheard the case and reversed, holding that partisan-gerrymandering claims were non-justiciable under the state constitution as well — a remarkable doctrinal reversal in less than a year. New York: Harkenrider v. Hochul (N.Y. Ct. App. 2022) struck down the Democratic-drawn congressional map, replaced it with a court-drawn map for 2022, and re-engaged in 2024 over a redrawing attempt. Wisconsin: Clarke v. Wisconsin Elections Commission (Wis. 2024) struck down the state-legislative maps as partisan gerrymanders under the Wisconsin Constitution. State courts have become the principal forum for partisan-gerrymandering claims after Rucho.

Allen v. Milligan (2023). The Supreme Court, 5-4, with Chief Justice Roberts writing for the majority and joined by Justice Brett Kavanaugh on key portions, held that Alabama's congressional map violated Section 2 of the Voting Rights Act because the state, with one majority-Black district out of seven and a 27% Black population, could and should have drawn a second majority-Black district. The decision rejected the State's argument that a more race-neutral map should be required and reaffirmed the Gingles framework. It was a significant reaffirmation of Section 2 vote-dilution doctrine — many observers in 2022 expected the Court to weaken Section 2; instead, Milligan strengthened it. Subsequent litigation in Louisiana, South Carolina, Georgia, and Texas worked out the implications.

Moore v. Harper (2023). The Court, 6-3, rejected the strongest version of the Independent State Legislature theory — the claim that under Article I, Section 4 (the Elections Clause), state legislatures have unreviewable authority over federal-election rules, including redistricting, free of state-court review under state constitutions. Moore held that state legislatures are subject to state-constitutional review. The decision preserved state-court avenues for federal-elections challenges; it would have been a transformative shift had the Court held the other way. Chief Justice Roberts wrote the majority opinion; Justices Clarence Thomas, Samuel Alito, and Neil Gorsuch dissented.

The composite legal map after these decisions: federal partisan-gerrymandering claims are dead (Rucho); state-constitutional partisan-gerrymandering claims are alive and being adjudicated in several states (PA, WI, NY); racial gerrymandering claims under the Equal Protection Clause are alive but doctrinally narrow (Shaw and progeny); Voting Rights Act Section 2 dilution claims are alive and stronger than many expected after Milligan; preclearance is dead (Shelby County); and the Independent State Legislature theory is dead in its strong form (Moore).

35.5 Measuring the bias: efficiency gap, partisan symmetry, compactness

Political scientists have developed a battery of quantitative measures of partisan bias in maps. Each measures something slightly different. Each has limits. None is the single right measure; a map evaluated against several measures gives a more reliable picture than a map evaluated against one.

Efficiency gap. Developed by Nicholas Stephanopoulos (law) and Eric McGhee (political science) in a 2014 paper, the efficiency gap measures the difference between the parties' "wasted" votes, divided by total votes. A "wasted" vote is one cast for a losing candidate (a "lost vote") or for a winning candidate beyond the 50%+1 needed for victory (a "surplus vote"). The intuition: gerrymanders waste the opposing party's votes through packing (lots of surplus votes in their packed districts) and cracking (lots of lost votes in their cracked districts). The bigger the asymmetry between the parties' wasted-vote rates, the bigger the efficiency gap. Stephanopoulos and McGhee proposed a threshold — efficiency gaps greater than about 7% in a state with eight or more districts — as suggestive of unconstitutional partisan gerrymandering. The measure was central to the Whitford v. Gill (later Gill v. Whitford) litigation that the Supreme Court ultimately disposed of on standing grounds in 2018, before Rucho finished off federal claims a year later. The efficiency gap is intuitive but has limitations: it implicitly assumes a roughly proportional baseline relationship between vote share and seat share, which the U.S. system does not have; it can move sharply with small changes in voter behavior; and it measures outcome bias, not intent.

Partisan symmetry. A maps shows partisan symmetry if a hypothetical 50%–50% statewide vote would produce a roughly 50%–50% seat outcome — that is, if the parties would win the same share of seats from the same share of votes. Partisan symmetry is the original political-science measure; Andrew Gelman and Gary King refined it in the late 1980s and 1990s. It requires inferring counterfactual results from observed results, usually through uniform partisan swing simulations. Maps with substantial partisan asymmetry — where one party would win, say, 60% of seats with 50% of the vote — are flagged.

Mean–median difference. Computed by taking the median district's vote share and subtracting the mean district's vote share. In a partisan-neutral map the two are close. In a packing-heavy gerrymander, the mean is dragged toward the packed districts (very high opposition vote shares) but the median is unaffected, producing a divergence.

Compactness. Compactness measures the geometric shape of districts. Various formulas: the Polsby–Popper measure (4π × area / perimeter²) — a perfect circle scores 1, and elongated or jagged districts score near 0. The Reock measure (district area / area of smallest enclosing circle). The Schwartzberg measure. None is definitive — many real-world geographic regions are not naturally compact, and the Voting Rights Act sometimes requires non-compact districts to ensure minority representation — but very low compactness scores often correlate with partisan or racial-gerrymandering concerns. The Princeton Gerrymandering Project publishes compactness data for every congressional map.

Simulation-based methods. A more recent approach is to generate thousands of computer-simulated maps using the same state's geography, the same population data, the same legal constraints (contiguity, equal population, VRA compliance), and to compare the actual enacted map against the distribution of simulated alternatives. If the enacted map is an outlier — falls in the most-pro-Republican 1% or most-pro-Democratic 1% of all simulated maps — that is evidence of partisan manipulation. Methods developed by Jonathan Mattingly (Duke), Wendy Tam Cho (Illinois), Kosuke Imai (Harvard), Devin Caughey (MIT), and others are now standard analytical tools. They were central to the League of Women Voters v. Commonwealth litigation in Pennsylvania (where 1,000 simulated maps were compared with the enacted map, which was an extreme outlier).

A composite finding from these measures across the 2010-cycle and 2020-cycle congressional maps: the 2010-cycle Republican gerrymanders in Wisconsin, North Carolina, Pennsylvania (before 2018 redrawing), Ohio, and Michigan each scored as outliers under multiple measures. Democratic gerrymanders in Maryland and Illinois scored as outliers as well, but the number of outlier maps was greater on the Republican side and the aggregate national effect — measured in expected House seats — was ten to fifteen seats in the Republicans' favor over the decade. In the 2020-cycle, the asymmetry narrowed: Republican aggressive maps in Texas, North Carolina (after the 2023 state-court reversal), Florida, Tennessee, and others; Democratic aggressive maps in Illinois, Maryland, Oregon, and (struck down) New York. Estimates of the net partisan bias of the 2020-cycle congressional maps vary by measure and by analyst; a fair summary is that the bias is smaller than the 2010-cycle bias but still tilted Republican by a few seats.

35.6 Geographic sorting: gerrymandering is not the only cause of safe seats

If you took every congressional district in the United States and ranked them by Cook Partisan Voting Index (PVI) — a measure of how each district has voted relative to the nation in recent presidential elections — you would find that, in 2024, only about thirty of 435 districts have PVIs in the genuinely competitive range (R+5 to D+5). The other 405 are "safe" by some margin. Many are extremely safe: D+30 urban districts, R+30 rural districts. The general election in those districts is essentially decided when the dominant party's primary closes.

How much of this lopsidedness is gerrymandering? Less than the popular understanding suggests, and that fact matters for reform.

The geographic sorting hypothesis. Jonathan Rodden of Stanford, in Why Cities Lose (2019) and in a series of academic papers with Jowei Chen and others, argues that the United States has experienced a profound spatial sorting of partisanship over the past several decades. Democrats are concentrated in urban centers, in inner-ring suburbs of major cities, and in college towns. Republicans are concentrated in exurbs, small cities, and rural counties. The sorting is dramatic: the median Democrat lives in a precinct where Democrats win by 60 percentage points; the median Republican lives in a precinct where Republicans win by 30 percentage points. (The asymmetry — Democrats more densely packed — is itself important.)

In a single-member-district plurality system, this geographic concentration automatically produces lopsided districts even with neutral lines. Computer simulations (Chen and Rodden 2013; many subsequent replications) show that any partition of, say, Pennsylvania into eighteen geographically compact, equal-population, contiguous districts — with no partisan-favoring or partisan-disfavoring rules — tends to produce a Republican-leaning delegation, because Democrats waste votes by piling them up in Philadelphia, Pittsburgh, and a handful of college towns. Estimates vary: Chen and Rodden's neutral-redistricting simulations of Pennsylvania in 2013 typically produced 9 or 10 Republican seats out of 18 even when the statewide vote was 50–50 or even slightly Democratic.

This finding has two important implications.

First, even fully fair, neutral, computer-drawn maps will produce a structural Republican advantage in the U.S. House under current geographic distributions of voters. The advantage is not as large as the 2010-cycle gerrymander advantage, but it is non-trivial. A reformer who promises that better maps will produce fully proportional outcomes is overpromising.

Second, the additional advantage produced by partisan gerrymandering is the marginal effect on top of the geographic-sorting baseline. Studies that compare actual enacted maps to neutral-rule simulations (the Mattingly–Imai approach) typically find that the 2010-cycle maps added 8–12 Republican seats above the geographic baseline; the 2020-cycle adds fewer.

The argument has critics. Bernard Grofman and others have argued that the geographic-sorting-only analysis underestimates how much line-drawing choices interact with geography — for example, where a state draws urban-suburban boundary lines can sharply change how many "blue" districts are created. Nicholas Stephanopoulos has argued that the comparison-to-simulations method itself depends on what rules the simulator uses, and that some rule choices (compactness, county integrity, VRA compliance) are themselves contested. Both criticisms are fair. The honest summary: geographic sorting matters; it is real; reform will not eliminate the structural single-member-district plurality bias; and on top of that, the marginal contribution of partisan line-drawing is real and consequential, but smaller than the popular discourse sometimes suggests.

35.7 Reform options: the menu and the trade-offs

A serious assessment of reform requires presenting each option's strongest case and its strongest critique. Six options recur in the policy literature.

Option 1: Independent commissions. Move the line-drawing authority from the legislature to an independent body. The Arizona, California, Michigan, and Colorado models. Strongest case: the empirical record from Arizona and California suggests that commission-drawn maps are more compact, produce more competitive seats in the marginal districts, and have less partisan bias than the legislature-drawn maps that preceded them. Independent commissions also produce more transparent processes — public hearings, published draft maps, comment periods. Strongest critique: commissions are not actually independent of politics — appointment processes can be captured, partisan affiliation rules can be evaded, and commissioners' implicit partisanship can shape outcomes despite formal independence. The California commission has produced maps that benefit Democrats in some cycles and the Arizona commission has been the subject of partisan disputes in every cycle. Commissioners are not specialists in line-drawing; they often defer heavily to professional staff and consultants, who themselves may have partisan affiliations. Critics on the right argue that "independent" reformers tend to be left-leaning (the foundations that fund commission ballot initiatives — the William and Flora Hewlett Foundation, the Schwarzenegger Institute, the League of Women Voters — are perceived as left-coded), and that the maps produced therefore have a subtle pro-Democratic bias. Whether this critique is correct is empirically contested; the published academic evidence does not support a strong claim that commission maps systematically favor Democrats more than legislature-drawn maps in similar states.

Option 2: Non-partisan staff drafting (Iowa model). Have a non-partisan legislative staff agency draft the map according to neutral criteria; have the legislature ratify or reject without amendment. Strongest case: Iowa's congressional and state-legislative maps over four decades have been compact, have low partisan-bias measures, and have produced competitive districts when geography permits. The system reduces the temptation to gerrymander by removing the partisan staff from the drafting process. Strongest critique: the model relies on the staff agency remaining genuinely non-partisan over time. In states with more polarized politics or more contested geography, the model would face stress that Iowa has not faced. The model also implicitly trusts the legislature to ratify even an outcome they dislike — Iowa's legislators have done this; legislators in other states might not.

Option 3: Algorithmic and computational redistricting. Use computer algorithms to generate maps that satisfy specified neutral criteria (compactness, equal population, contiguity, VRA compliance) and select among them by transparent rules. The Imai/Caughey approaches; the Princeton Gerrymandering Project's analytical tools; various academic proposals. Strongest case: algorithms can generate thousands of options at low cost, can be transparent about the criteria used, and can take human partisanship out of the optimization step. Strongest critique: every algorithm requires human-chosen rules. What weight to put on compactness versus county integrity versus VRA compliance is itself a choice; the rule-choosers can game the rules to favor preferred outcomes. Computer-drawn does not mean neutral. There is also a participatory critique: redistricting is a political question, and removing it from politicians removes it from the people who elected those politicians, which raises legitimacy questions of its own.

Option 4: Multi-member districts with proportional voting. Replace single-member-plurality elections with multi-member-district proportional or semi-proportional systems (cumulative voting, ranked-choice in multi-member districts, party-list proportional). Most other democracies use some such system. Strongest case: in proportional systems, the seat distribution closely tracks the vote distribution, and the geographic-sorting problem largely disappears. Gerrymandering becomes much less consequential because outcomes depend primarily on vote totals, not on lines. Strongest critique: a federal change to multi-member districts would require Congress to repeal the Uniform Congressional District Act of 1967, which mandated single-member districts. Multi-member districts have their own pathologies — they can dilute racial-minority representation if not designed carefully, they may weaken local representation, and they require voter familiarity with new ballot mechanisms. The reform is also difficult politically: incumbents in safe districts have weak incentives to support a system that would put them at risk. Chapter 39 examines proportional-representation systems in cross-national perspective.

Option 5: Federal anti-gerrymandering legislation. Congress could legislate national redistricting standards under Article I, Section 4 (the Elections Clause). Several proposals: the John R. Lewis Voting Rights Advancement Act would have updated the Voting Rights Act preclearance formula; the Freedom to Vote Act would have set national standards including independent-commission requirements, partisan-fairness criteria, and bans on extreme partisan gerrymanders. Both bills passed the House in 2021 but failed to overcome Senate filibusters in 2022. Strongest case: a federal standard would address the inconsistency across states, in which voters in Iowa get a non-partisan process and voters in North Carolina or Maryland get aggressive gerrymanders. Federal action would also reach states whose courts have refused to police partisan gerrymandering (as North Carolina's did after 2023). Strongest critique: a federal statute would require either (a) a Senate filibuster carve-out (sixty votes are unlikely on this issue), (b) a constitutional crisis if signed and challenged, or (c) bipartisan support that is not now apparent. Even if enacted, federal standards would face Article I Section 4 challenges and would require an enforcement mechanism — the Department of Justice, the courts, or both — that would have to be staffed and funded across multiple administrations. The federalism critique is also serious: redistricting has been a state function since the founding, and federal rules raise legitimate Tenth Amendment concerns.

Option 6: Constitutional amendment. Several amendment proposals would establish national redistricting principles or create federal oversight machinery. Strongest case: an amendment would settle the question definitively and would not depend on shifting Senate majorities. Strongest critique: the amendment process requires two-thirds of each house of Congress and three-fourths of the state legislatures. State legislatures are precisely the bodies whose redistricting authority would be limited, making ratification extraordinarily unlikely. Amendments are the longest possible road; for a politically polarized issue, often an impossible road.

A further reform-design observation: the existing reform movement has produced more change in the blue and purple states (California, Michigan, Colorado, Virginia) than in the red states. Independent commissions established by ballot initiative require ballot-initiative procedures — and not every state has them. Of the twelve states that have adopted some form of commission for congressional redistricting, eleven have ballot-initiative or constitutional-amendment-by-initiative procedures available. The exception, Virginia, used a legislative process. The reform geography therefore tilts toward states where the political culture supported initiatives in the first place. This has produced an asymmetric reform landscape: the states with the most aggressive gerrymanders are often the ones least reachable by reform.

35.8 The "competitive elections" question — does the cure work?

A common reform argument is: better maps will produce more competitive districts, which will produce representatives who must appeal to the median voter, which will reduce polarization. The argument has intuitive force. It also has serious complications.

The conservative critique. More-competitive districts may produce more polarized representatives, not less. The argument: in a safe district, the incumbent does not need to fear a primary challenge and can govern toward the median of the district; in a competitive district, the incumbent must satisfy the activist base in their primary or face being replaced by a more-extreme candidate, and then they must perform polarized signals to mobilize their base in the general. Some empirical evidence supports this critique: studies of congressional behavior in the 2010s found that members in safe seats had voting records closer to their districts' median voters than members in competitive seats had to theirs. The argument is associated with John Sides, Lynn Vavreck, and political scientists at the AEI–Brookings Election Reform working group, and with portions of the Cato Institute's analysis.

The progressive counter. Safe districts produce primary-driven extremism more reliably than competitive ones. In a safe Republican district, the only election that matters is the Republican primary, and the activist base of that primary is to the right of the district median; the same logic applies on the Democratic side. The result is "primary by ideologue, govern by ideologue." Competitive districts force at least some appeal to the general-election median. Empirical evidence is mixed; studies by Eric McGhee, Seth Hill, Boris Shor, and Nolan McCarty have produced different conclusions in different specifications.

A more careful synthesis. The relationship between district competitiveness and representative ideology is heterogeneous. It depends on the structure of the primary (open versus closed), on the density of activist organization in the district, on whether one party dominates the partisan landscape (a competitive district is meaningfully different in a 50–50 state than in a 70–30 state), and on whether the incumbent is well-funded. Reformers who promise that competitive districts will produce moderate representatives are overstating; defenders of safe districts who claim that competitive seats produce extremists are also overstating. The institutional design question is: do you want representatives accountable to the median primary voter (safe seats) or to the median general-election voter (competitive seats)? Reasonable people disagree. The empirical effects of districting reform on polarization are smaller than partisans on either side claim.

35.9 What this chapter does not settle

Three things this chapter has not tried to do.

It has not declared partisan gerrymandering unconstitutional. Rucho declared it non-justiciable in federal court, and that is the operating legal reality for the 2020s. Whether it is normatively incompatible with democracy depends on a theory of democracy. Robert Post and Reva Siegel have argued that gerrymandering breaches the principle of political equality; Justin Levitt has argued that it corrodes the consent on which democratic legitimacy depends; Adrian Vermeule has counter-argued that aggressive partisan competition is itself a feature of democratic politics, and that elite "good-government" reformers should not be entrusted to override it. The chapter has presented the strongest version of each argument and asked you to take the question seriously rather than predetermining the answer.

It has not endorsed any specific reform. Each option has merits and costs. Independent commissions look better empirically than legislature-drawn maps but are not a panacea; algorithmic redistricting is transparent but rule-dependent; multi-member districts solve some problems and create others. A reader with strongly held priors will find arguments in this chapter for and against each option. That is the design.

It has not predicted the next decade. The 2030 redistricting cycle will follow the 2030 Census, which will be conducted under whichever administration is in power and under whichever Census protocols emerge. State legislative composition in 2030 will determine who controls map-drawing in most states. State-court composition will determine what challenges succeed. The federal Voting Rights Act will continue to constrain — or, if Congress acts (one direction or the other), be modified. The reform organizations now active — Common Cause, the Princeton Gerrymandering Project, RepresentUs, the Brennan Center, the Center for Election Innovation, the Cato Institute, the Ethics and Public Policy Center — will continue their work. What that work produces depends on millions of choices made by voters, legislators, and judges between now and 2031.

35.10 The recent state-level battles: 2020-cycle litigation in detail

The legal framework described in Section 35.4 is not abstract; it is being applied, right now, in courtrooms across the country, with real consequences for House delegations and state legislatures. A tour of the 2020-cycle litigation that has actually moved seats clarifies how the doctrines work in practice.

Alabama (Allen v. Milligan, 2023). Alabama's seven-seat congressional map, drawn by the Republican legislature after the 2020 Census, contained one majority-Black district (the 7th) and six majority-white districts. Black Alabamians make up roughly 27% of the state's voting-age population. Plaintiffs argued under Section 2 of the Voting Rights Act that the map diluted Black voting power and that a second majority-Black district was both possible and required. A three-judge district court agreed, ordering the map redrawn. The Supreme Court initially stayed that order in early 2022 (allowing the original map to be used in the 2022 elections), then in June 2023 affirmed the lower court 5–4. Chief Justice Roberts and Justice Kavanaugh joined the three liberal justices in the majority. Alabama's redrawn map for 2024, drawn after the legislature's first attempt was rejected by the federal court as still non-compliant, produced two majority-Black districts; both elected Black Democrats. The decision triggered immediate parallel litigation in Louisiana (similar facts, similar outcome — a court-ordered second majority-Black district), South Carolina (Alexander v. NAACP, 2024, where the Court ruled differently on a racial-gerrymandering claim, allowing the existing map to stand), and Georgia (Section 2 litigation ongoing).

Louisiana. A federal district court applying the Allen framework ordered Louisiana to redraw its six-seat congressional map to include a second majority-Black district. The Republican legislature, after appellate skirmishing, complied with a map containing two majority-Black districts (the 2nd and the new 6th). The redrawn map was used in 2024 and produced a second Black Democratic representative.

North Carolina. The trajectory in North Carolina is the most dramatic doctrinal reversal of the cycle. In 2022, the state supreme court (then with a 4–3 Democratic majority) struck down the Republican-drawn congressional map as a partisan gerrymander violating the state constitution (Harper v. Hall). A court-substituted map was used in 2022, producing a 7–7 split delegation. The November 2022 election flipped the state supreme court's partisan composition to 5–2 Republican. The new majority granted rehearing in Harper and, in 2023, reversed: partisan-gerrymandering claims, the new majority held, are non-justiciable under the North Carolina Constitution. The Republican legislature redrew the map in 2023 to be more aggressively Republican; the 2024 elections under the new map produced a 10–4 Republican delegation. North Carolina was therefore the rare case in which one state's congressional map was redrawn three times in a single decade — once by the legislature, once by the courts, once by the legislature again — and the partisan composition of the state supreme court determined which map operated when. Moore v. Harper arose from this same litigation; the U.S. Supreme Court rejected the strong Independent State Legislature theory but did not reach the partisan-gerrymandering question, which the state court had already mooted by reversing course.

New York. The 2022 Democratic-drawn congressional map was struck down by the New York Court of Appeals (the state's highest court) as violating both the state constitution's partisan-gerrymandering prohibition and the procedural requirements added by the 2014 redistricting amendments. A court-appointed special master drew the 2022 map, which produced a more balanced delegation. In 2024, the state's Independent Redistricting Commission re-engaged, deadlocked, and the legislature drew a slightly more Democratic-favorable map for the 2024 cycle, which was upheld against legal challenge as not crossing the partisan-gerrymander threshold. Democrats gained seats in New York in 2024 partly as a result; the state's congressional delegation moved from a 15–11 Democratic edge in 2022 to a 19–7 edge in 2024. New York is the case study for Democrats gerrymander when they can, and state courts can constrain them. (Case Study 02 covers New York in detail.)

Wisconsin. The state-legislative maps drawn by Republicans after 2010 — Act 43, the Wisconsin map at issue in Gill v. Whitford — were extreme by every measure: in 2018, Democrats won a majority of the statewide legislative vote and 36% of state assembly seats. After Rucho foreclosed federal partisan-gerrymandering claims, litigation moved to the state courts. The April 2023 election flipped the Wisconsin Supreme Court to a 4–3 liberal majority. In Clarke v. Wisconsin Elections Commission (December 2023), the new majority struck down the state-legislative maps under the Wisconsin Constitution. The legislature redrew the maps in 2024 under Governor Tony Evers's terms. The 2024 state-legislative elections under the new maps produced a Democratic gain of 14 seats and a much more competitive legislative landscape. Wisconsin demonstrates: after Rucho, state-court composition matters enormously, and state-court partisan flips can fundamentally reshape redistricting law within a state in months rather than decades.

Michigan. Michigan's Independent Citizens Redistricting Commission, established by 2018 ballot initiative, drew the state's 2022 congressional and state-legislative maps. The commission's process was protracted and produced maps that drew criticism from both parties. The maps performed roughly proportionally in 2022 and 2024: Democrats won a slight majority of statewide votes and a slight majority of seats. Compactness scores improved relative to the prior decade's legislature-drawn maps. The Michigan commission has subsequently faced legal challenges (some state-legislative districts were struck down for race-related concerns under Shaw doctrine in 2024) but the structural shift to citizen control has held.

Texas, Florida, Tennessee, Ohio, and others. Republican-controlled legislatures in these states drew maps in 2021–22 that consolidated Republican advantage. Texas added two seats from reapportionment and drew them as Republican-leaning; Florida's governor pushed an aggressive map that eliminated a Black-plurality district (litigation ongoing under Section 2); Tennessee cracked Nashville across three districts; Ohio's map, repeatedly struck down by the Ohio Supreme Court as violating the state's 2018 redistricting reform, was used anyway after a federal court allowed the unconstitutional map to be used because no constitutional alternative had been finalized in time. Ohio is the case of a state where the redistricting reform — a bipartisan commission with strict criteria adopted by 75% of voters in 2018 — failed in implementation, because the relevant actors (the Republican-dominated Redistricting Commission) refused to comply with the rules even after being told four times by the state supreme court to do so. Reform-on-paper is not reform-in-practice when the political will to enforce is absent.

Maryland and Illinois. Both Democratic-controlled states drew aggressive Democratic maps. Maryland's 2021 map was struck down by a state court as a violation of the state constitution; the legislature redrew it less aggressively. Illinois's map survived legal challenge and produced a 14–3 Democratic delegation in a state where Democrats win statewide elections by single digits. Illinois is the case study for how, in the absence of state-court hostility to partisan gerrymandering, a unified Democratic state government can draw a map that delivers a wildly disproportional delegation.

A net assessment of the 2020-cycle: the partisan-bias of congressional maps, measured aggregate across all states, is smaller than 2010-cycle bias but still tilted Republican by approximately three to six seats nationally. The bias narrowed because: (a) Section 2 litigation produced new majority-Black districts in Alabama and Louisiana, both adding Democratic seats; (b) New York's struck-down Democratic map was redrawn favorable to Democrats again in 2024; (c) Michigan moved to a commission that produced more proportional outcomes; (d) state courts in Pennsylvania, Wisconsin, and (briefly) North Carolina constrained Republican aggressive maps. The bias remains because: (a) Texas, Florida, and post-2023 North Carolina drew aggressive Republican maps; (b) Ohio operated for the cycle on maps the state supreme court repeatedly held unconstitutional; (c) geographic sorting continues to advantage Republicans even under neutral maps. The aggregate national picture is that both parties gerrymandered when they had unified state-government control and weak state-court constraints, that the number of opportunities tilted Republican (more Republican-controlled state governments than Democratic-controlled), and that Allen v. Milligan produced a counter-pressure on the Republican side that did not exist in the 2010 cycle.

35.11 The 2010 cycle in retrospect: REDMAP and its consequences

To understand the 2020-cycle litigation, you have to understand the 2010 cycle that produced it. Case Study 01 covers REDMAP in detail; this section sketches the strategic logic and the empirical effects.

The Republican State Leadership Committee (RSLC), a 527 organization founded in 2002, launched the Redistricting Majority Project (REDMAP) in 2009 with an explicit two-step strategy: first, target-spend in 2010 state-legislative races to capture state legislatures in advance of the post-2010 redistricting cycle; second, draw congressional and state-legislative maps that would lock in Republican advantage for the decade. The 2010 election — a wave year for Republicans nationally, on top of REDMAP's targeted state-level investments — delivered both halves. Republicans gained roughly 700 state-legislative seats in 2010 and took unified control of redistricting in states including Pennsylvania, Ohio, Michigan, Wisconsin, North Carolina, and Florida. The party then drew maps in those states that, by every standard measurement, produced an unusually large partisan advantage in the U.S. House for most of the decade.

The strategic insight of REDMAP was that, dollar-for-dollar, state-legislative races were profoundly underpriced in 2010. A state senate seat could be flipped with $50,000–$200,000 in targeted spending; a U.S. House seat would cost millions. But state-legislative seats, in redistricting years, controlled the lines for ten years' worth of U.S. House delegations. The leverage was extraordinary. RSLC documented its own spending: roughly $30 million across the 2010 cycle, with claimed effects in 100+ state-legislative races. David Daley's Ratfked (2016) details the program from RSLC's internal perspective and is the most-cited journalistic account.

The empirical effects of the 2010-cycle maps on the U.S. House: estimates from various studies converge on a Republican advantage of approximately 10–15 House seats above the geographic-sorting baseline through 2018. Brennan Center analyses, Princeton Gerrymandering Project simulations, and academic papers (Stephanopoulos and Warshaw 2017; Chen and Cottrell 2016) all produced estimates in this range. The 2018 elections, fought partly on already-redrawn maps (Pennsylvania's after the League of Women Voters decision; Florida's after League of Women Voters of Florida v. Detzner in 2015), narrowed but did not eliminate the bias. The 2020 cycle, with state courts more active, with Allen v. Milligan requiring new majority-Black districts, with Democratic states drawing their own aggressive maps, produced a smaller but still-real Republican structural advantage in the House.

A crucial fact that the popular discourse sometimes obscures: the 2010-cycle gerrymanders were the most aggressive single-cycle gerrymanders in modern history. They were not normal. The aggressive maps drawn by Democrats in Maryland and Illinois in the same cycle were, by quantitative measures, less extreme than the most-aggressive Republican maps. This is not a partisan judgment; it is a measurement. The efficiency-gap data, the partisan-symmetry data, the simulation-comparison data, and the compactness data all rank the 2010-cycle Pennsylvania, Wisconsin, North Carolina, and Ohio Republican maps as outliers even compared to other gerrymandered maps in U.S. history. This empirical fact has to coexist with the equally true empirical fact that Democrats also gerrymandered when they could (Maryland 2010, Illinois 2010, New York 2022). Both facts are true. The asymmetry was real; the symmetry-of-incentive (both parties gerrymander when given the chance) was also real.

35.12 The Cook PVI distribution: a snapshot of safe and competitive seats

The Cook Partisan Voting Index (PVI), developed by Charlie Cook of the Cook Political Report and now updated by David Wasserman, scores each congressional district by its presidential vote relative to the national average across the two most-recent presidential elections. A district that voted for the Republican presidential candidate by 5 percentage points more than the nation overall is R+5; a district that voted for the Democratic presidential candidate by 10 points more than the nation overall is D+10. The PVI is a useful summary of underlying partisan lean, distinct from the candidate-quality factors that determine any single House race.

The 2024 PVI distribution, computed from the 2020 and 2024 presidential results across the new district lines:

• About 30 districts have PVIs in the genuinely competitive range (R+5 to D+5). These are the districts where a competent candidate from either party could win in a normal year.
• About 80 districts have PVIs in the lean range (R+6 to R+10, or D+6 to D+10). Generally winnable by the favored party but losable in wave years.
• About 150 districts have PVIs in the likely range (R+11 to R+20, or D+11 to D+20). The dominant party usually wins comfortably.
• About 175 districts have PVIs of R+21 or D+21 or stronger. Effectively decided in the dominant-party primary.

The distribution has shifted over time. In 1996, 164 districts had Cook PVIs of R+5 to D+5. By 2024, that number was approximately 30 — an 80% decline in the count of genuinely competitive districts over twenty-eight years. The decline is the central institutional fact of contemporary U.S. House politics. Whether the decline is primarily driven by gerrymandering, by geographic sorting, by ideological sorting (Democrats and Republicans becoming more ideologically distinct so that districts that lean one way lean firmly one way), or by some combination is a question political scientists actively debate. The honest answer is that all three factors contribute, in different proportions in different states. Gerrymandering is part of the picture but not the whole picture.

The implication for governance is profound. When 405 of 435 House members face general elections that are essentially decided, their primary electoral concern is the dominant-party primary, and they govern accordingly. The median primary voter in a competitive primary is more ideologically extreme than the median general-election voter in the same district, because primary turnout is lower and skewed toward activists. The result is what political scientists call primary-driven polarization: representatives respond to the median primary voter, who is more partisan than the median general-election voter, and Congress polarizes accordingly. Gerrymandering reform is sometimes pitched as a polarization remedy on this logic; Section 35.8 above examined the conservative critique of that argument.

35.13 The reform asymmetry: why blue and purple states reform faster than red ones

A pattern emerges from the post-2010 reform record that deserves explicit attention: the states that have moved most decisively toward independent commissions, neutral-criteria drafting, or other anti-gerrymandering arrangements are predominantly blue or purple states (California, Michigan, Colorado, Virginia, New York's Independent Redistricting Commission), while the states that retain aggressive legislature-drawn maps with the most-extreme partisan biases are predominantly red states (Texas, Florida, Tennessee, post-2023 North Carolina, and Ohio in implementation). The asymmetry is not accidental, and it has at least three causes worth naming honestly.

First, the ballot-initiative geography. Most successful redistricting reforms have come through ballot initiatives — citizen-initiated constitutional amendments that bypass legislatures hostile to reform. Of the 24 U.S. states that allow citizen-initiated constitutional amendments or statutes for redistricting reform, roughly two-thirds have used the procedure to enact some form of reform. But many red states (Tennessee, Texas, Georgia, Alabama, South Carolina, Louisiana) do not allow citizen-initiated constitutional amendments at all, foreclosing the most effective reform path. A reform movement aimed at all fifty states cannot rely on a procedure that exists in only some states. Federal action would be required to reach the others; federal action is, as Section 35.7 noted, politically unavailable.

Second, the reform-coalition composition. The organizations driving redistricting reform — Common Cause, the League of Women Voters, RepresentUs, the Brennan Center, the Princeton Gerrymandering Project, the Hewlett Foundation, Schwarzenegger Institute — are coded by friend and foe alike as left-leaning or, at minimum, as drawing more support from left-leaning donors than from right-leaning ones. This perception (whether or not it is empirically accurate about the maps the reforms produce) creates resistance in conservative state legislatures and conservative voter populations. A serious reform movement needs visible right-of-center allies, and the Cato Institute, the R Street Institute, and select state-level conservative reform organizations have provided some — but the reform coalition's center of gravity is left of center, and that fact shapes which states embrace it.

Third, the strategic-incumbency calculation. A state legislature that controls redistricting and is currently producing maps favorable to its dominant party has no incentive to give up that authority. Republican legislators in Texas have nothing to gain from a commission that would produce more competitive seats; Democratic legislators in Illinois face the same incentive. The commissions that have been adopted are mostly the result of either (a) ballot-initiative bypass of the legislature, or (b) legislature concession in moments of unusual cross-pressure. The asymmetry in opportunity (ballot initiatives in some states but not others) compounds with the asymmetry in coalition strength (left-coded reform organizations) to produce the asymmetry in outcomes (more reform in blue and purple states than in red ones).

The honest summary: reform proceeds where the procedural and political conditions allow, and those conditions are not evenly distributed across states. The most-aggressive gerrymanders are often in states least reachable by reform. Federal action would address this, but federal action requires Senate supermajorities or filibuster reform that is unlikely in the near term. The reform movement therefore proceeds piecemeal, achieving real changes in California, Michigan, Colorado, and Virginia while making little progress in Texas, Florida, or Tennessee. Whether that piecemeal progress is enough to address the structural problem, or whether structural change requires federal action that is politically unavailable, is a question the reform movement has not resolved.

35.14 Connections and what comes next

Chapter 36 takes the next analytical step: from the drawing of districts to the administration of elections within them, and to the contemporary debate over voting rights. Voting access, voter identification, mail balloting, early voting, registration rules — all are state-level choices with measurable empirical effects on turnout. The Voting Rights Act of 1965, the Shelby County v. Holder (2013) decision dismantling preclearance, the wave of state voting laws enacted after 2020, and the 2024 election experience all sit there.

Chapter 37 turns from these specific institutional pathologies to the broader question of democratic erosion: how the institutional pathologies covered in Chapters 34, 35, and 36 — money in politics, gerrymandering, voting access — combine with the polarization documented in Chapter 25 and the constitutional hardball documented across the institutions chapters to put the system as a whole under stress. Chapter 38 closes the book with a sober assessment of what reforms are possible, what reforms are likely, and what is genuinely up to the next generation of voters and legislators to choose.

A final note. Two anchor projects of this textbook converge in this chapter. The Democracy Audit asks you to examine your own congressional district. The exercises ask you to find your district's map, look at its compactness measures, trace its history of redrawing, identify the political-control conditions under which it was drawn, and ask honestly: when you cast your vote, are you choosing your representative, or is the map choosing for you? The honest answer in many districts is both, and in proportions that depend on facts about your state. That is the analysis the chapter has been building. Politicians have, in important measure, chosen their voters. Voters can, with attention and effort, push back. The how is the rest of this section of the book.