Read the Report

Executive Summary

key finding

Despite a record turnout in the 2020 general election, only 10% of eligible Americans nationwide cast ballots in primary elections that effectively decided the winners in a supermajority (83%) of Congressional seats.

Partisan primaries are the “Primary Problem” in our politics today.

  • Due to geographic self-sorting and partisan gerrymandering, 83% of congressional districts lean so Democratic or so Republican (“safe”) that the only election of consequence is the primary election.⁸
  • As a result, in partisan primaries, a small minority of voters decide the vast majority of congressional elections — fueling political polarization and preventing problem-solving.  

Partisan primaries disenfranchise voters.

  • In 2020, of the 361 “safe” congressional districts, 151 of the districts had no competition in the dominant-party primary, denying any voters a say in the outcome.
  • In partisan primaries in the remaining 210 “safe” districts, most voters in the non-dominant party effectively had no voice in choosing their representative.
  • And in 10 states, nearly 11 million independent voters were prohibited from participating in either party’s primaries altogether.
  • In total, only 23 million of America’s 235 million voters (10%) effectively elected 83% of Congress.

Partisan primaries distort representation.

  • Voters who participate in primary elections are often unrepresentative of both their own party, and especially the electorate as a whole, producing similarly unrepresentative outcomes in the candidates they elect.
  • New polling data from Colorado’s 3rd Congressional District, for example, found that the Republican primary electorate that voted for challenger Lauren Boebert over incumbent Rep. Scott Tipton was nearly twice as likely (60%) to identify as “very conservative” compared to general election voters (25%).

Partisan primaries fuel political division & dysfunction.

  • With no electoral competition in the general election, most members of Congress’ only threat to reelection is the potential of being “primaried” by someone to their ideological extreme.
  • Research demonstrates that incumbents alter their behavior to appeal to primary voters — including which voters they are incentivized to represent and how they vote. Meanwhile, other research finds that potentially more moderate candidates opt out of running in the first place because they know they cannot win.

Nonpartisan primaries can solve the Primary Problem.

  • States can adopt nonpartisan primaries — as already used in California, Washington, and Nebraska, and most recently adopted in Alaska — that allow all voters to cast a ballot in a single primary with all candidates on the same ballot. The top finishers advance to the general election, where whoever earns majority support wins.
  • Nonpartisan primaries give every voter an equal voice, have higher voter participation rates, produce more representative outcomes, and improve governing incentives by ensuring elected leaders are accountable to a broader swath of the electorate. One recent study found that, among new members of Congress elected between 2003-2018, those elected in top-two nonpartisan primaries were more than 18 percentage points less extreme than those elected in closed partisan primaries.
  • Alaska’s new “Final-Four Voting” system combines a top-four nonpartisan primary with ranked choice voting general elections — an improvement over top-two primaries that helps increase competition and levels the playing field for candidates outside of both major parties by eliminating the “spoiler” effect.


On the morning of January 6th, 2021, newly elected Congresswoman Lauren Boebert (R-CO-3) tweeted, “Today is 1776,” and egged on the mob that arrived in Washington, DC to “stop the steal” of the presidential election.

On the National Mall, former President Trump fired up the crowd:  “You have to get your people to fight. And if they don’t fight, we have to primary the hell out of the ones that don’t fight. We primary them.”

Primary (n): a preliminary election to nominate the candidates of each political party and/or to reduce the field of all candidates ahead of the general election.

Primary (v):
to challenge an incumbent politician from within their own party, usually from the far-left or far-right, for falling out of step with their base.

Despite the violent insurrection in the afternoon, 147 Republican members of Congress voted to object to the Electoral College vote later that same evening — many fearing their next party primary.

Rep. Boebert cast one of those votes to overturn the election, and she knows a thing or two about primaries; in 2020, she was one of just eight challengers (five Republicans and three Democrats) who defeated an incumbent in a primary election.

Rep. Boebert’s election in Colorado’s 3rd District, as this paper will elaborate, presents an illustrative case study in how partisan primaries have become the Primary Problem in our politics today — fueling political division and dysfunction that not only undermines good governance and problem solving but now threatens democracy itself.

That the 116th Congress had a near record-low 15% approval rating, yet 95% of its members were reelected in the last election, underscores how the lack of desired representation and results in government is a systems problem, not unique to any person or party.

The problem is not just who we elect, but rather how we elect. In particular, how few Americans actually do the electing, in partisan primaries.

This paper analyzes the 2020 primaries and offers three major findings: first, a small minority of voters effectively decide the vast majority of who ends up in Congress in partisan primaries; second, the disproportionately partisan primary electorate incentivizes leaders appeal to the extremes and change their behavior; third, nonpartisan primaries can solve the Primary Problem and foster better governing incentives by ensuring our leaders are accountable to all of the voters in their district.

i. The data in this paper includes all elections for the U.S. House in 2020. It does not include elections for the U.S. Senate.

1 “Congress and the Public,” Gallup (December 2020).

2 “Election Results, 2020: Incumbent Win Rates by State,” Ballotpedia (February 2021).