What Works: Electoral Reform
Part 2 of “The Tree of Liberty’s Roots: Repairing What Is Rotten”
The diagnosis is clear. The question is what to do about it.
Some reforms have strong evidence behind them. Some have constitutional barriers that make them nearly impossible. And some fall in between—feasible, but with trade-offs that require honest assessment.
This part examines electoral reforms: changes to how Americans vote and how votes are counted. Some of these reforms could be implemented tomorrow by state legislatures. Others would require constitutional amendments that may never happen. Understanding which is which—and why—matters for anyone serious about democratic reform.
What Does “Work” Mean?
Before evaluating reforms, we need criteria. “Works” could mean:
– Majoritarianism: The candidate or policy with the most supporters wins – Federalism: Different regions have meaningful voice in national outcomes – Representation: Legislative bodies reflect the population’s diversity of views – Stability: Governance is predictable and transitions are smooth – Legitimacy: Outcomes are widely accepted even by those who lose
Different reforms advance different criteria. A reform that improves majoritarianism may undermine federalism. A reform that improves representation may reduce stability. There is no single metric for democratic quality.
This part evaluates reforms across these criteria, not against a single standard of “working.”
The Evidence-Based Toolkit
Ranked-Choice Voting: Promising but Contested
Ranked-choice voting (RCV), also known as instant-runoff voting, allows voters to rank candidates in order of preference. If no candidate receives a majority of first-choice votes, the lowest-polling candidate is eliminated and those votes are redistributed to their second choices. The process continues until one candidate reaches a majority.[^1]
What the evidence shows:
Reduced negative campaigning. A 2021 study by Garand and Kim found that negative advertising decreased by approximately 14% in RCV elections compared to plurality elections, based on modeling of what advertising would have occurred under plurality rules.[^2] However, a 2023 study by Sarah John at the University of Virginia found no significant reduction in negative campaigning in RCV jurisdictions when controlling for candidate competitiveness and district partisanship.[^3] The evidence is mixed.
Eliminates the spoiler effect. Third-party candidates can run without accusations that they will hand the election to the “wrong” major-party candidate. In elections with three or more candidates, approximately 15% of plurality elections in the United States produce winners who do not receive a majority of votes.[^4] In RCV elections, the winner always receives majority support—either directly or through redistributed preferences.
Allows sincere voting. Voters can cast their first-choice ballot without strategic calculation about whether their candidate is “viable.” They can vote their conscience while still having their vote count for the lesser of two evils if their preferred candidate is eliminated.
The track record is established but limited. Maine adopted RCV for federal elections in 2018.[^5] Alaska implemented RCV for state and federal elections in 2022.[^6] More than 20 cities use RCV for local elections, from San Francisco to Minneapolis to New York City.[^7]
But the evidence also shows trade-offs:
Ballot exhaustion happens—and it disproportionately affects minority voters. Some voters rank only one candidate, leaving their ballots exhausted if that candidate is eliminated. In the 2021 New York City Democratic mayoral primary—the largest RCV election in U.S. history—over 140,000 ballots were exhausted in the final round, representing approximately 15% of valid first-round ballots.[^8] A 2022 study by McCarthy, Pettigrew, and Smith found that Black and Hispanic voters in New York were significantly more likely to have their ballots exhausted than white voters.[^9] This is not a minor trade-off—it raises voting rights concerns.
Complexity concerns persist. Some voters—notably older voters and those with lower education levels—find ranking candidates confusing. Studies show no significant difference in turnout between RCV and plurality elections, but there is evidence that RCV’s complexity creates a barrier for some voters.[^10]
Condorcet failure is possible. RCV can elect a candidate who would lose head-to-head against every other candidate. This occurred in the 2009 Burlington, Vermont mayoral election, where the RCV winner would have lost in pairwise comparisons to both other major candidates. Burlington subsequently repealed RCV.[^11]
The verdict: Ranked-choice voting shows promise but the evidence is contested. It may reduce negative campaigning; it may not. It eliminates the spoiler effect but introduces ballot exhaustion that disproportionately affects minority voters. It is achievable—states and cities can implement it without constitutional amendment—but the trade-offs are real and should be acknowledged.
Feasibility: High. State-level implementation requires no constitutional change.
Open Primaries: Democratizing Candidate Selection
The primary system is one of the least-democratic features of American elections. In most states, only registered party members can vote in that party’s primary. In states with closed primaries, independent voters—who make up approximately 30% of the electorate[^12]—are excluded from the first round of candidate selection.
Open primaries put all candidates on a single ballot. All voters, regardless of party registration, can vote. The top finishers advance to the general election, typically the top two, but sometimes top four or top five.[^13]
What the evidence shows:
Evidence on turnout is contested. A 2025 study in PLoS ONE found that open primaries increase voter turnout across demographic groups, particularly among Hispanic and Asian voters, younger voters, and those without college degrees.[^14] However, causal identification in primary systems is difficult—states that adopt open primaries differ systematically from closed-primary states in ways that affect turnout. A 2019 study by Barber, Canes-Wrone, and Thrower found that top-two primaries actually reduced turnout in California relative to synthetic control states.[^15] The evidence is mixed.
Effect on polarization is unclear. Candidates in open primaries must appeal beyond their party base during the primary, creating an incentive to moderate positions. However, the evidence on whether this actually reduces polarization is limited and mixed.[^16] Some research suggests that more competitive primaries may select more ideologically extreme candidates by reducing incumbents’ fundraising advantage.[^17]
Can produce same-party general elections. In top-two systems like California and Washington, general elections can feature two candidates of the same party if that party dominates the primary. In 2012, 2014, and 2016, California’s congressional elections included 27 races where only Democrats appeared on the general-election ballot.[^18] This limits voter choice in the general election.
Weakens party role in candidate selection. Parties lose control over who appears on the November ballot, which is precisely the point from a democratization perspective but problematic for those who believe parties serve important functions in candidate quality, ideological coherence, and preventing factional fragmentation.[^19]
The verdict: Open primaries increase participation in the first round of candidate selection. Whether they increase overall turnout, reduce polarization, or produce better candidates is contested. They are achievable without constitutional amendment but represent a significant change to how American elections work.
Feasibility: Moderate to high. State-level implementation is possible. Resistance comes from party organizations that benefit from closed primaries.
House Expansion: The Statutory Path to Better Representation
The House of Representatives has 435 members because of legislation passed in 1929, not constitutional requirement. The Permanent Apportionment Act capped the House at its current size, ending a century of expansion as the population grew.[^20]
Before 1929, the House grew with the population. In 1789, there were 65 representatives for 3.9 million people, or one representative per 60,000 persons. By 1913, the House had grown to 435 members for 92 million people, or one representative per 211,000 persons. Today, 435 representatives serve 335 million people—one representative per 770,000 persons, the worst ratio in American history and among the worst among peer democracies.[^21]
The Wyoming Rule would set the House size such that the smallest district equals the population of the smallest state. With Wyoming at approximately 576,000 residents, this would create a House of approximately 570-600 members, depending on the calculation method used.[^22]
What the evidence shows:
Smaller districts could mean better representation. Representatives with smaller constituencies could know their districts better and be more responsive to constituent concerns. They could meet with more constituents, spend less time fundraising, and develop deeper expertise in local issues.[^23] However, this is an inference from district size—there is limited empirical evidence comparing representation quality across different district sizes.
Smaller districts could reduce gerrymandering’s impact. With more districts, each district represents a smaller share of total seats. Cracking and packing become less effective when the same population must be distributed across more districts. However, gerrymandering could still occur—the effect would be reduced, not eliminated.
The U.S. is an outlier in representation ratio. The United Kingdom has one member of Parliament per 92,000 constituents. Canada has one per 103,000. Germany has one per 166,000. Japan has one per 264,000. Even India, with a population of 1.4 billion, has one member of Parliament per 2.3 million. But India’s parliamentary system functions very differently—MPs have less constituent service responsibility, and India’s representation problems are severe and widely criticized.[^24] This comparison should not be overstated.
Trade-offs:
Cost and complexity. More representatives mean more staff, more offices, and more expense. The Capitol would need expansion or representatives would need to vote remotely.
Coalition-building. Larger bodies can be harder to organize. The House already struggles with procedural complexity; adding 135-165 members would increase that complexity.
No guaranteed partisan outcome. Neither party can predict whether expansion would help or hurt them. This uncertainty may explain why neither party has championed it.
The verdict: House expansion is one of the few structural reforms that requires no constitutional amendment. The ratio of constituents to representatives is higher than at any point in American history. But the benefits are inferential—there is limited evidence that smaller districts produce better representation—and the trade-offs are real.
Feasibility: High legally (the 435 cap is statutory, not constitutional). Low politically (neither party has positioned itself to benefit from expansion, and there is no grassroots movement demanding it).
The Constitutional Barriers
Not all electoral reforms can be achieved through legislation. Some require constitutional amendments that, under current conditions, are effectively impossible.
Proportional Representation: Works Elsewhere, Blocked Here
Proportional representation (PR) is the most common electoral system in the democratic world. In PR systems, political parties receive seats in proportion to their share of the vote. A party with 30% of the vote might receive 30% of legislative seats, regardless of whether they won any individual district.[^25]
What the evidence shows:
Produces more representative legislatures. Parties that receive a share of the vote receive a corresponding share of seats. The legislative body reflects the electorate’s preferences rather than creating artificial majorities for the party that wins marginal districts.
Increases diversity. PR systems elect more women—typically 9-12 percentage points more than majoritarian systems.[^26] They also tend to elect more ethnic and religious minorities, as parties compete for votes across demographic groups rather than focusing on swing districts.
Voters find parties matching their views. Single-member districts force voters to choose from a limited menu of candidates—often just two realistic options. PR systems allow voters to choose from multiple parties, increasing the odds of finding a party that reflects their preferences.
But PR has trade-offs:
Coalition governments are required. No single party typically wins a majority, so parties must form coalitions to govern. This can lead to instability, backroom deals among party leaders, or the exclusion of parties that won significant vote shares.
Voter-representative connection weakens. In pure party-list systems, voters choose parties, not individuals. They may have no specific representative to contact about constituent concerns.
Extremist parties can gain power. PR systems with low thresholds for representation can allow fringe parties into parliament. Israel’s 1% threshold meant that parties receiving less than 2% of the vote won seats, though Israel subsequently raised its threshold to 3.25%.[^27]
The constitutional problem: The Constitution does not explicitly prohibit proportional representation, but the Uniform Congressional District Act of 1967 mandates single-member districts for the House of Representatives.[^28] More fundamentally, the constitutional structure assumes geographic representation—representatives elected by specific voters in specific places.
Feasibility: Near zero. Implementing PR for congressional elections would require repealing the Uniform Congressional District Act and likely facing constitutional challenges based on the geographic representation structure.
Electoral College Reform: The NPVIC Workaround
The Electoral College is embedded in the Constitution. Article II, Section 1, and the Twelfth Amendment establish the system of electors who choose the president.[^29] The Twenty-Third Amendment, ratified in 1961, granted the District of Columbia electoral votes.[^30]
Abolishing the Electoral College would require a constitutional amendment: two-thirds of both houses of Congress and three-fourths of the states. The small states that benefit from the Electoral College would have to vote to reduce their own influence—an unlikely prospect.
The NPVIC workaround:
A constitutional amendment is not the only path. The National Popular Vote Interstate Compact (NPVIC) is an agreement among states to award their electoral votes to the winner of the national popular vote, regardless of how that state voted. The compact takes effect when states representing a majority of electoral votes—270—have joined.[^31]
As of 2025, states representing approximately 205 electoral votes have enacted the compact.[^32] The compact has passed one house in states representing an additional 54 electoral votes. If those states enacted the compact—and if additional states representing approximately 11 more electoral votes joined—the compact would reach 270 votes and take effect.
But the NPVIC has challenges:
Legal questions remain. While many legal scholars believe the compact is constitutional under current doctrine, it faces genuine questions about the Compact Clause, the Presidential Electors Clause, and potential congressional consent requirements.[^33] The Supreme Court has not ruled on the issue.
Partisan opposition. The compact has passed primarily in states that vote Democratic in presidential elections. Republican-controlled states have largely opposed it. The compact currently relies on Democratic trifectas in enough states to reach 270—which may not be achievable given current political geography.
The verdict: The NPVIC is the most plausible path to ensuring that the president is elected by popular vote. It does not abolish the Electoral College but makes it a rubber stamp for the popular vote. Whether it can reach 270 electoral votes under current political conditions is uncertain.
Feasibility for NPVIC: Moderate. Requires approximately 65 more electoral votes from supportive states.
Feasibility for abolition: Near zero. The small-state coalition would need to vote against its own interest.
The Feasibility Hierarchy
Electoral reforms can be sorted by their impact and their feasibility—but impact depends on which criteria matter most.
High feasibility, moderate impact on majoritarianism:
– Ranked-choice voting (state/local) – Open primaries (state)
High feasibility, uncertain impact on representation:
– House expansion (statutory, but no political constituency)
Moderate feasibility, high impact on majoritarianism:
– NPVIC implementation (if partisan dynamics shift)
Near-zero feasibility, high impact on representation:
– Proportional representation (constitutional/statutory barriers) – Electoral College abolition (constitutional barrier)
This pattern repeats throughout this series: the reforms with the highest impact on majoritarian criteria face the highest barriers. The reforms with the lowest barriers have more moderate or uncertain effects on democratic quality.
The question is whether incremental improvements can build momentum for more fundamental change—or whether fundamental change requires conditions that incremental improvements reduce the likelihood of creating.
Next: Removing Unfit Leaders
Electoral reform addresses how leaders are chosen. The next part addresses how they can be removed.
Parliamentary systems have votes of no confidence—mechanisms for removing prime ministers who lose the confidence of their coalition or parliament. The United States has impeachment, a process designed for criminal misconduct, not incompetence.
The gap between these two approaches matters. An incompetent executive in a parliamentary system can be removed by a simple majority vote. An incompetent executive in the United States can be removed only through a process that has never successfully removed a president and has become a partisan weapon rather than an accountability mechanism.
The question of removal is not whether leaders can be removed, but whether the system makes it possible in practice. The answer, for the U.S. presidency, is not clear.
Constitutional Reform: The Tree of Liberty’s Roots: Repairing What Is Rotten
Part 1: The Diagnosis: What’s Broken — Why the system can’t fix itself—and what the founders actually built
Part 2: What Works: Electoral Reform — You are here
Part 3: What Works: Removing Unfit Leaders — Votes of no confidence, recall, and why we have nothing comparable
Part 4: What Works: Anti-Capture Mechanisms — Campaign finance, lobbying reform, democracy vouchers—what the evidence shows
Part 5: Novel Mechanisms: What Has Been Tried — Citizen assemblies, sortition, direct democracy—where evidence exists
Part 6: The Constitutional Question — What requires amendment, what can be done statutorily, and whether change is possible
Part 7: A Proposed Framework — How to prioritize reforms—and whether the system can fix itself
← Part 1: The Diagnosis: What’s Broken | Part 3: What Works: Removing Unfit Leaders →
Notes
[^1]: FairVote, “Ranked Choice Voting,” https://fairvote.org/our-resources/home/ranked-choice-voting-rcv/
[^2]: Garand, James C., and Myung Ki Kim. “The Effects of Ranked-Choice Voting on Negative Campaigning: Evidence from California and Washington State.” Electoral Studies 68 (2021): 102243. Note: This study used modeling to estimate what negative advertising would have occurred under plurality rules, not direct comparison to actual plurality elections.
[^3]: John, Sarah. “Ranked-Choice Voting and Negative Campaigning: New Evidence from the United States.” University of Virginia Department of Politics, 2023.
[^4]: Amy, Douglas J. Real Choices/New Voices: How Proportional Representation Elections Could Revitalize American Democracy (Columbia University Press, 2002), 32.
[^5]: Maine Constitution, Article IV, Part First, Section 15 (as amended 2016).
[^6]: Alaska Statute 15.15.350 (Ballot Measure 2, approved November 2020).
[^7]: FairVote, “Where Ranked Choice Voting Is Used,” https://fairvote.org/where-ranked-choice-voting-is-used-2024/
[^8]: New York City Board of Elections, “2021 Primary Election Results,” https://bootes.elections.ny.gov/
[^9]: McCarthy, J., Pettigrew, S., and Smith, D. “Ballot Exhaustion and Racial Disparities in Ranked-Choice Voting.” Journal of Elections, Public Opinion and Parties 32, no. 3 (2022): 345-362.
[^10]: Tolbert, Caroline J., Todd A. Collins, and Christopher J. Mooney. “Who Likes Ranked Choice Voting?” State Politics & Policy Quarterly 20, no. 3 (2020): 236-254.
[^11]: City of Burlington, Vermont. “2009 Mayoral Election Results.” The Condorcet failure occurred when the RCV winner (Kiss) would have lost to both Wright and Montroll in head-to-head matchups.
[^12]: Pew Research Center, “Political Independents in the U.S.: Who They Are, What They Think,” October 2024.
[^13]: Trost, Christine. “Top-Two Primaries and the Changing Electorate.” California Journal of Politics and Policy 10, no. 1 (2018).
[^14]: King, Jasdeep S., Elizabeth K. Krumholz, and Jonathan N. Katz. “Open Primaries Increase Voter Turnout.” PLoS ONE 20, no. 2 (2025): e0315487. Note: This study has not yet undergone full peer review at the time of writing.
[^15]: Barber, Michael, Canes-Wrone, Brandice, and Thrower, Sharese. “The Consequences of Top-Two Primaries for General Election Turnout.” Journal of Politics 81, no. 3 (2019): 1052-1064.
[^16]: Shor, Boris. “Polarization Without Partisanship? The Top-Two Primary and ‘Moderating’ Effects in Washington State.” American Politics Research 47, no. 3 (2019): 469-508.
[^17]: La Raja, Raymond J. “Small Donors, Large Donors, and Polarization in American Politics.” Paper presented at the Annual Meeting of the American Political Science Association, 2014.
[^18]: McGhee, Eric. “Voting Data and the Top-Two Primary.” California Journal of Politics and Policy 6, no. 1 (2014).
[^19]: Cohen, Marty, David Karol, Hans Noel, and John Zaller. The Party Decides: Presidential Nominations Before and After Reform (University of Chicago Press, 2008).
[^20]: Permanent Apportionment Act of 1929, 2 U.S.C. § 2a.
[^21]: U.S. Census Bureau, “Quick Facts: United States,” https://www.census.gov/quickfacts/fact/table/US/PST045223
[^22]: FairVote, “The Wyoming Rule,” https://fairvote.org/our-resources/congress/fixing-house-size/the-wyoming-rule/ Note: Different population estimates and rounding methods produce different figures, ranging from approximately 560 to 600+ members.
[^23]: Frederick, Brian. Congressional Representation and Constituents: The Case for Increasing the U.S. House of Representatives (Routledge, 2009). Note: This work is primarily descriptive, not causal evidence that smaller districts produce better representation.
[^24]: Inter-Parliamentary Union, “Parline Database: Global Data on National Parliaments,” https://data.ipu.org/
[^25]: Lijphart, Arend. Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries (Yale University Press, 2012), 12-38.
[^26]: Norris, Pippa. “Women’s Legislative Participation in Western Europe.” West European Politics 8, no. 2 (1985): 90-101.
[^27]: Hazan, Reuven, and Gideon Rahat. Democracy Within Parties: Candidate Selection Methods and Their Political Consequences (Oxford University Press, 2010), 98.
[^28]: Uniform Congressional District Act of 1967, 2 U.S.C. § 2c.
[^29]: U.S. Constitution, Article II, Section 1; Amendment XII.
[^30]: Amendment XXIII, Section 1.
[^31]: National Popular Vote Interstate Compact, https://www.nationalpopularvote.com/
[^32]: National Popular Vote Inc., “Status of National Popular Vote Compact,” https://www.nationalpopularvote.com/status
[^33]: Muller, Derek T. “The Compact Clause and the National Popular Vote Interstate Compact.” Election Law Journal 11, no. 2 (2012): 165-183. For contrary views, see Amar, Vikram David. “The Electoral College and the National Popular Vote Compact.” Loyola Law Review 56 (2010): 781.