In most U.S. congressional districts, the general election is a formality. The real decision is made months earlier, by a small fraction of voters most Americans never think about.
This single structural fact — that primary elections, not general elections, determine most congressional outcomes — explains more about why Congress behaves the way it does than almost any other variable in American politics. It explains the polarization. It explains the fundraising obsession. It explains why compromise has become so institutionally dangerous. And it explains why electing "better people" to Congress, absent structural reform, tends to produce the same behaviors.
If you want to understand American politics, understand the primary.
The Math That Nobody Talks About
Of the 435 House seats, roughly 350-380 are competitive in only one direction. They're either safely Democratic or safely Republican, meaning the winner of the majority party's primary will win the general election as a near-certainty.
In those districts, the primary electorate — the voters who actually decide who represents the community — is a small, unrepresentative slice of the district's population. Turnout in congressional primaries averages around 15-20% of registered voters in the party. In many races, it's lower.
Who votes in primaries? Consistently, across decades of political science research, the answer is: the most ideologically committed, the most engaged with party politics, and the most activated by high-stakes framing. In other words, the voters most likely to have been reached by exactly the kind of fear-based, existential messaging described in the fundraising industrial complex.
"This single structural fact — that primary elections, not general elections, determine most congressional outcomes — explains more about why Congress behaves the way it does than almost any other variable in American politics."
The result is a selection mechanism that systematically elevates candidates who appeal to the most activated partisan base rather than the broadest cross-section of the district. A moderate Democrat who can win a general election in a swing district may not survive a primary if the primary electorate is dominated by voters who view compromise as betrayal. A Republican who has built a reputation for bipartisan deal-making may find that reputation is a liability, not an asset, when it comes to the 15% of registered Republicans who decide primaries.
The Structural Incentive to Extremism
Once elected, members face the same constraint in reverse: any behavior that might be used against them in a primary is institutionally dangerous.
Voting with the other party — even on issues where your constituents have cross-partisan views — hands a primary opponent a weapon. Being seen as too friendly with the opposition, too willing to negotiate, too interested in actual governance — all of these become attack lines. The fundraising email writes itself: "Representative X voted with [enemy party] 23 times last year. Is that who you want representing you?"
This is not a hypothetical. It is a well-documented pattern of institutional behavior. Members who built careers on bipartisan collaboration — Richard Lugar, Bob Bennett, Lisa Murkowski (who lost a primary, then won on a write-in campaign) — have faced primary challenges precisely because their collaborative records were framed as disloyalty.
The rational response to this environment, for a member who wants to keep their job, is to avoid any behavior that could be characterized as crossing the line. The definition of "crossing the line" is set by the primary electorate, which is more extreme than the general electorate, which is more extreme than the district as a whole.
This is how a representative democracy ends up being governed by the preferences of a small minority of the most ideologically activated voters in each party. Not through any malicious intent. Through the arithmetic of primary elections.
The Safe Seat Paradox
There is a paradox embedded in this structure that is worth naming explicitly.
Safe seats — districts so reliably partisan that the outcome is never really in question — sound like a feature. They provide stability. They allow long-serving members to develop expertise. They let members take principled stands without constant electoral threat.
"The safer the general election, the more dangerous the primary. The more dangerous the primary, the more extreme the behavior incentivized."
In practice, safe seats intensify the primary dynamic rather than alleviating it. A member in a safe seat has only one real electoral threat: a primary challenge from someone more ideologically pure. The general election is irrelevant. Which means the only voters who matter — the only voters whose preferences the member needs to track — are the primary electorate.
The safer the general election, the more dangerous the primary. The more dangerous the primary, the more extreme the behavior incentivized. This is the safe seat paradox: the districts that should produce the most secure, most independent-minded legislators instead produce some of the most behaviorally constrained.
The Reform Toolkit
There are structural interventions that could change these dynamics, and several are being implemented at the state level with measurable results.
Ranked-choice voting in primaries allows voters to express preferences across multiple candidates without spoiler effects, which tends to produce winners with broader appeal. Alaska adopted ranked-choice voting in 2020 and immediately saw it produce an outcome that the old system couldn't: a moderate Republican winning in a primary over several more extreme candidates because she was the second choice of enough voters to build a majority.
This is the third installment of "Power, Plainly Explained," a series examining the machinery of American government. Next: "What Happens When AI Enters Campaign Strategy."