The Math: 158 Million Votes, 44,000 Decided It
In the 2020 U.S. presidential election, Joe Biden received 81.2 million votes nationally and Donald Trump received 74.2 million — a difference of 7 million votes. Biden won the Electoral College 306-232. None of that national margin mattered. The outcome was determined entirely by three states: Arizona (Biden by 10,457 votes), Georgia (by 11,779), Wisconsin (by 20,682). A combined shift of approximately 43,000 votes — 0.027% of total ballots cast — would have thrown the election to the House of Representatives.
This is not anomalous. The 2000 election was decided by 537 votes in Florida. The 2004 election turned on 118,000 votes in Ohio. The 2016 election turned on 77,744 votes across Michigan, Pennsylvania, and Wisconsin — combined. The structural reason for this pattern is the Electoral College combined with winner-take-all state allocation: 48 states give all electoral votes to the statewide plurality winner regardless of margin, meaning votes for the minority candidate in any state generate zero electoral outcome.
The practical consequence: presidential campaigns are not national campaigns. In 2020, over 96% of general election campaign spending was concentrated in 12 states. If you lived in California, Texas, New York, Illinois, or any state with a predictable partisan lean — you effectively did not exist as a presidential voter. No candidate visited you, advertised to you, or needed your vote.
The Senate: The Most Malapportioned Chamber in the Democratic World
Every state receives exactly two U.S. Senators, regardless of population. This was a deliberate compromise at the Constitutional Convention in 1787, designed to secure ratification from small states. In 1790, the population ratio between the largest state (Virginia, ~748,000) and the smallest (Delaware, ~59,000) was approximately 13 to 1. The same structural rule applied to 2020's ratio of California (39.5M) to Wyoming (576,000) produces a disparity of approximately 68 to 1.
In practical terms: a Wyoming voter's Senate representation is worth roughly 68 times a California voter's Senate representation. A senator from Wyoming, representing 576,000 people, has exactly one vote — the same vote as a senator from California, representing 39.5 million people. The Senate's 100 members collectively represent states holding 100% of the population, but the 52 senators from the 26 smallest states represent just 17% of the U.S. population.
| State | Population (2020) | Senators | People per Senator | Relative Weight |
|---|---|---|---|---|
| Wyoming | 576,000 | 2 | 288,000 | 68x California |
| Vermont | 643,000 | 2 | 321,000 | 61x California |
| Alaska | 733,000 | 2 | 366,000 | 54x California |
| North Dakota | 779,000 | 2 | 390,000 | 50x California |
| Texas | 29.1M | 2 | 14,550,000 | 0.5x California |
| California | 39.5M | 2 | 19,750,000 | Baseline |
This structure has direct policy consequences. The filibuster — which requires 60 Senate votes to end debate — means that senators representing as few as 10% of the U.S. population can block legislation indefinitely. Voting rights legislation that passed the House with broad support has died in the Senate multiple times in the past decade, unable to clear the 60-vote threshold, with opposition coming substantially from senators representing a small fraction of the national population.
Source: U.S. Census Bureau 2020 Decennial Census; Senate vote records, 117th-118th Congress; Congressional Research Service — Senate filibuster rulesGerrymandering: Engineering Safe Seats
Congressional district lines are drawn by state legislatures after each decennial census. The party controlling the legislature controls the map. Gerrymandering works through two documented techniques. Packing concentrates opposition voters into a small number of districts they win by overwhelming margins, wasting their surplus votes. Cracking divides opposition voters across multiple districts where they form a minority in each. The combined effect: one party can win a majority of seats while receiving a minority of total votes.
The Supreme Court addressed partisan gerrymandering directly in Rucho v. Common Cause (2019). The 5-4 majority opinion, written by Chief Justice Roberts, held that federal courts cannot adjudicate claims of partisan gerrymandering because the Constitution provides no "judicially discoverable and manageable standards" for evaluating them. The decision effectively removed federal judicial oversight of partisan map-drawing — leaving the practice subject only to state courts (under state constitutions) and Congress (which has not passed nationwide redistricting reform).
The consequence is documented in the Cook Political Report's annual competitive seat analysis: in a typical congressional election year, approximately 35 of 435 House seats are genuinely competitive. The remaining 400 are safe enough that the general election is a formality — the effective election is the party primary.
Congressional primaries — the elections that actually choose most members of Congress — typically draw 10–15% of eligible adult voters. In safe districts, whoever wins the primary wins the seat. The median primary voter is consistently more ideologically extreme than the median general election voter and far more extreme than the median eligible adult. Roughly 10–15% of the eligible population is effectively selecting the House of Representatives.
First-Past-the-Post and the Two-Party Lock
The United States uses plurality voting — the candidate with the most votes wins, even without a majority. This produces what political scientists have called Duverger's Law: plurality systems strongly tend toward two-party competition. The mechanism is the spoiler effect: a third-party candidate ideologically similar to a major-party candidate drains votes without accumulating enough to win, handing victory to the more distant candidate. Voters who understand this dynamic vote strategically for the lesser of two evils rather than their preferred candidate. Over time, third parties are squeezed out.
The United States has had two dominant parties continuously since 1856. No third-party candidate has won a presidential election since Abraham Lincoln's Republican Party in 1860 — itself only possible because it displaced the collapsed Whig Party. Third parties have won presidential electoral votes three times since then (1912, 1924, 1968) and influenced outcomes several times without winning. None has built durable representation.
Alternative systems address the spoiler problem differently. Ranked-choice voting (RCV) lets voters rank candidates in order of preference; if no candidate wins an outright majority, the last-place candidate is eliminated and those ballots are redistributed to voters' next choice. No spoiler effect exists. Proportional representation allocates legislative seats based on vote share — a party receiving 15% of the vote gets approximately 15% of seats. Both systems are used in democracies with substantially higher turnout and broader ideological representation than the U.S.
Voter Turnout: Most Recent National Election
Source: IDEA Voter Turnout Database 2024; U.S. ranks 31st of 50 established democracies in presidential election turnoutMoney After Citizens United
Citizens United v. FEC (2010) held that the First Amendment prohibits government from restricting independent political expenditures by corporations, associations, or labor unions. The immediate structural consequence: Super PACs — political action committees that can raise unlimited funds from any source and spend unlimited amounts on political advertising, provided they don't formally coordinate with a candidate's campaign. The financial effect was immediate and sustained.
Beyond Super PACs, Citizens United enabled the expansion of 501(c)(4) "social welfare" organizations — nonprofits that can engage in political activity as long as it is not their "primary purpose." These organizations are not required to disclose their donors publicly, producing what campaign finance researchers call "dark money": political spending whose source is unknown to voters. OpenSecrets estimates dark money spending at approximately $1 billion in the 2020 cycle and significant portions of the 2022 and 2024 cycles, though by definition the full accounting is not possible.
“The appearance of corruption is not enough to justify restricting political speech. We now return control of our political discourse to its rightful owners — the American people.”
Justice Kennedy — Citizens United v. FEC majority opinion (2010) — The American people to whom control was returned include corporations, unlimited Super PACs, and anonymous 501(c)(4) donors.The Reform Landscape
Several structural reforms have been proposed or partially implemented. The National Popular Vote Interstate Compact (NPVIC) is an agreement among states to award all their electoral votes to the winner of the national popular vote — but only once states totaling 270 electoral votes have joined. As of early 2026, 17 states plus the District of Columbia have joined, representing 209 electoral votes. The Compact is 61 electoral votes short of triggering. It has passed in several additional state chambers without becoming law.
Ranked-choice voting has been adopted in Maine (federal elections, 2016), Alaska (2020), and several dozen cities. Australia has used it for the House of Representatives since 1918. Ireland uses a version called Single Transferable Vote for its parliament. No U.S. state has adopted proportional representation for federal elections.
Independent redistricting commissions — bodies that remove map-drawing authority from state legislatures — have been adopted in California, Arizona, Michigan, Colorado, Virginia, and several other states after ballot initiative campaigns. Their record is mixed: some have produced demonstrably more competitive maps; others have been challenged and modified by state courts or legislatures. The fundamental problem — that whoever draws the lines has power — is not eliminated, only redirected.
The For the People Act (H.R. 1), passed by the House in 2019 and 2021, would have established national independent redistricting commissions, automatic voter registration, nationwide early voting requirements, and limits on certain campaign finance practices. It passed the House both times. It died in the Senate both times, unable to break a filibuster — the same structural barrier whose math is described in the Senate section above.
The International Comparison
The United States ranks 31st out of 50 established democracies in presidential election turnout, per the International Institute for Democracy and Electoral Assistance (IDEA) database. This is not primarily a story of voter apathy — it reflects structural choices embedded in U.S. electoral administration that most peer democracies do not share.
Automatic voter registration exists in all 50 U.S. states in some form, but implementation varies enormously. Voter registration in the U.S. requires individual action; in most European democracies it is automatic upon reaching adulthood or becoming a citizen. The U.S. holds most major elections on Tuesdays, a choice dating to 1845 and designed around the travel schedules of agricultural workers. Most high-turnout democracies hold elections on weekends or declare Election Day a public holiday. Australia and Belgium make voting compulsory, with small fines for non-compliance; both have turnout rates above 85%.
Proportional representation — in which a party's share of seats matches its share of votes — is used in Germany, Netherlands, Sweden, Norway, Denmark, New Zealand, and most other parliamentary democracies. It structurally prevents the seat-vote divergence documented in U.S. gerrymandering cases. It also tends to produce multi-party legislatures rather than two-party systems, which changes the incentive structure around legislative compromise. The U.S. uses single-member district plurality voting for both the House and the Senate — a combination that political scientists consistently identify as the strongest structural driver of two-party polarization.
The filibuster — which requires 60 Senate votes to advance most legislation — is not constitutional. It is a Senate rule that emerged from a procedural accident in 1806. No equivalent exists in any other established democracy. Its practical effect is that 41 senators representing states with approximately 11% of the U.S. population can block any legislation. The populations that minority can represent are mathematically predetermined by the Senate's equal-state structure: low-population states, which trend Republican, benefit disproportionately from both malapportionment and filibuster calculus.
What the Princeton Study Found
In 2014, political scientists Martin Gilens (Princeton) and Benjamin Page (Northwestern) published "Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens" in Perspectives on Politics. The study analyzed 1,779 policy outcomes between 1981 and 2002 against the stated policy preferences of average citizens, economic elites, and organized interest groups.
Their findings: average citizen preferences had near-zero independent impact on policy outcomes. When controlling for elite preferences, mass opinion "adds little or nothing" to the predictive model. Economic elites and organized interest groups — particularly business groups — had substantial independent impact. The paper concluded that U.S. governance "in the face of opposition by economically-elite Americans and large-scale business-oriented interest groups... rarely prevails."
The study has been critiqued on methodological grounds — primarily that distinguishing elite preferences from average citizen preferences is difficult when they often overlap, and that the historical period studied may not reflect current conditions. The authors have responded to these critiques in subsequent work. The paper remains the most-cited quantitative study of whose preferences drive U.S. policy outcomes, and its findings are consistent with the structural record documented throughout this investigation.
The Electoral College concentrates presidential elections into a handful of states. Senate malapportionment gives small-state voters up to 68x the Senate representation of large-state voters. Partisan gerrymandering — now unreviewable by federal courts after Rucho — produces legislatures where seat share routinely diverges from vote share by 15–25 percentage points. Outside political spending has grown 13x since Citizens United. Primary elections drawing 10-15% of eligible adults select most of Congress. The Princeton study found average citizen preferences had near-zero independent impact on policy outcomes. These are not perceptions of dysfunction — they are the documented outputs of the system as designed and as modified by law and court decision.