Executive summary

Eight states hold primary runoff elections if no candidate wins a majority of votes in a party primary: Alabama, Arkansas, Georgia, Louisiana, Mississippi, Oklahoma, South Carolina, and Texas. Two states hold primary runoff elections if no candidate reaches a lower share of votes: North Carolina (30%) and South Dakota (35%).

FairVote has studied all federal primary runoff elections held since 1994. Below are our key findings.

• Runoffs bring dramatic declines in voter turnout: Dating back to 1994, 97% of congressional primary runoffs have had lower turnout than the initial primary election.
• It’s getting worse: The 2024 primary season had the largest drop in turnout since FairVote began tracking runoffs in 1994, with a turnout decline in every single runoff and a median of 63% fewer voters participating in the runoff than in the initial primary.
• Runoffs are wildly expensive: In 2024, federal primary runoffs cost taxpayers at least $6.9 million, and likely $9-12 million total (this total estimate excludes costs for runoffs in Mississippi). These costs did not include any statewide races; a Kennesaw State University study found that the 2020 statewide U.S. Senate runoff in Georgia cost the state a whopping $75 million.
• Most runoff winners get fewer votes than they did on Election Day: 81% of candidates who won runoffs in 2024 received fewer votes in the runoff than they did in their initial primary, essentially negating the runoff’s purpose. Additionally, every 2024 primary runoff was decided by a margin smaller than the number of first-round voters who chose to stay home.
• Runoffs can intensify negative campaigning, as demonstrated by a case study in Texas.
• Ranked choice voting solves the runoff problem: Ranked choice voting works like an instant runoff, capturing the majority’s preferences in a single election when turnout is highest and the electorate best reflects the public. Research shows that RCV also reduces negative campaigning.

Introduction

Runoff elections have been employed by U.S. states since the nineteenth century. Runoffs are also fairly common abroad; 12 out of 18 Latin American nations use a form of runoffs, as do European nations like Finland and France.

Today, 10 states hold primary runoff elections for the U.S. Senate and U.S. House of Representatives. In these states, if no candidate surpasses a certain threshold (typically 50% of the vote), the top two vote-getters compete in a delayed runoff election to determine the party’s nominee. (South Dakota and North Carolina have different thresholds, only holding a runoff if no candidate achieves 35% or 30%, respectively.)

In this report, we focus on federal primary runoff elections. Over the past 30 years, states have held 292 federal primary runoff elections. During that time, 126 of the primary runoff winners were ultimately elected to serve in Congress.

The main objective of the runoff system is to identify a nominee with majority support. Runoffs are far from representative, though. Turnout almost always declines in runoff elections, and the voters who participate tend to be older, Whiter, and wealthier than those who voted in the first round. This undermines the goal of identifying a broadly supported nominee. Moreover, administrators have to hold another taxpayer-funded election, and candidates and voters have to endure a prolonged campaign season.

FairVote has tracked these dynamics in all federal primary runoff elections since 1994. This report documents our findings and highlights how issues with runoffs affected the 2024 elections. We also show how ranked choice voting — also known as “instant runoff voting” — offers a better, faster, cheaper alternative.

Record-low voter turnout in 2024 runoffs

This runoff season saw an unprecedented 63% median decline in turnout between the initial primary and the runoff. This is a greater decline than any year since we began tracking in 1994.

Almost two-thirds of primary voters failed to vote in the 16 runoffs conducted across six states this year — a significant increase from 2022’s median turnout decline of 52%. This appears to be part of a larger trend in which the turnout gap between primaries and primary runoffs is growing.

Indeed, the turnout decline in every 2024 primary runoff was brutally high. Whereas the occasional runoff in prior years saw a turnout increase, no 2024 primary runoff was spared a turnout decline. Of all 292 runoffs we’ve tracked since 1994, seven of the 30 lowest-turnout runoffs (in comparison to their respective primaries) occurred this year, with three runoffs (in MS-2, TX-29, and TX-35) losing over 80% of primary voters. Most 2024 runoffs saw turnout decline far greater than the 30-year median of 41%.

81% of winning candidates won with fewer votes than they received in the first round

When a candidate wins a runoff election with fewer votes than they earned in the initial primary, it negates the purpose of the runoff. In other words, runoff elections are resulting in fewer voters choosing nominees — not more.

In 2024, 13 out of 16 runoff winners (81%) received fewer votes in the runoff than in the first round. As turnout declines have gotten worse over the last two decades, this problem has become more frequent. In fact, 2024 was the fifth consecutive cycle in which the majority of runoff winners received fewer votes than they earned in the initial primary.

Additionally, over the last two decades, 33% of primary runoff winners were different from the leader in the initial primary, and that number jumped to 50% in 2024. These “come-from-behind” wins are a natural feature of runoff elections, but they come into question when we consider the large number of runoffs with extremely low turnout.

Runoff turnout decline is greater in communities of color

Turnout decline tends to be higher among voters of color, which may discourage candidates from appealing to those communities in runoffs and reduce attention on the issues impacting them.

Turnout decline in 2024 runoffs was more significant among voters of color, whose turnout dropped by a median of 70%, compared to White voters at 64%. The disparity may be a result of obstacles to voting that disproportionately impact voters of color. Ultimately, runoffs lead to shallower representation of the electorate.

Every 2024 runoff election lost more voters than the margin between finalists

When turnout declines, and a candidate wins the runoff with fewer votes than they earned in the first round, it is difficult to know whether the runoff winner is truly the party’s preferred candidate. Every 2024 primary runoff had a margin of victory smaller than its turnout decline. In other words, the number of voters who participated in the first round but not the runoff was large enough to have changed the runoff result.

For example, in the Republican runoff in Georgia’s 2nd District, Brian Jack beat Mike Dugan by 11,961 votes in the runoff, but 23,067 fewer people voted than in the initial primary. If those voters had returned to the polls to vote in the runoff, they could have easily changed the result. Of course, it is possible that some of these voters stayed home because they didn’t have a preference between the two finalists. However, it’s also possible that many of them did have a preference, and did not vote because of the inconvenience of returning to the polls or lack of awareness that another election was being held.

Some runoffs were decided by paper-thin margins, which makes it harder to claim the result was representative. Runoffs in Texas’s 7th and 29th Districts were both decided by fewer than 100 votes, while the runoff in TX-35 was decided by just five votes. In fact, the typical runoff winner in 2024 would have lost their seat if just 13% of “missing” primary voters had shown up to vote for their competitor.

These findings demonstrate that primary runoff winners aren’t necessarily representative of the full electorate, but just of those who are able to show up twice.

Runoffs burned at least $7 million of taxpayer money

In 2024, taxpayers spent a minimum of $6.9 million holding primary runoff elections, and an estimated total of $9-12 million. We obtained costs and made estimates based on records obtained from five out of the six states that held federal primary runoff elections this year. Our data is available publicly at this link, including a description of our methodology for estimating costs for counties that did not provide them.

Note that these costs did not include any statewide races, and included only 16 U.S. House races – representing about 20% of the total of 84 districts across the states that held runoffs this year. When Georgia held a runoff election for its U.S. Senate seats in 2020, it cost the state $75 million. This year, some states (including Georgia) did not conduct any runoffs in their largest counties.

In states with runoffs, voters are not getting “bang for their buck” when it comes to election costs. Voters are spending millions of dollars on runoffs that do not deliver meaningful majorities. Moreover, the expense can be made redundant by ranked choice voting, as discussed later.

The more time between primary and runoff, the lower the turnout

In an effort to give overseas and military voters enough time to request and return runoff ballots, several states have opted to increase the number of days between primary elections and the subsequent runoffs. However, extending the gap between elections also correlates with a drastic decline in voter turnout.

Since 1994, runoffs held within 30 days of the initial primary had a median turnout decline of 33%, while those with a gap greater than 30 days had a median decline of 48%. The marginal turnout decline for each day between primaries and runoffs (0.2%) shows that moving runoffs to a later date, even by a week, can potentially damage voter turnout.

Case study: “Runoffs are just brutal” in Texas’s 23rd District

Runoff elections can create additional negative campaigning, because they give the leading campaigns extra time to target their single strongest opponent.

For example, take the Republican primary in Texas’s 23rd Congressional District. In the first round, incumbent Representative Tony Gonzales faced attacks from three challengers who cast him as a “Republican in name only” and as “corrupt.” All three challengers agreed that ousting Gonzales motivated them just as much as representing their constituents.

When internet personality Brandon Herrera advanced to the runoff to compete head-to-head with Gonzales, the runoff ultimately became a proxy war for competing Republican factions; one U.S. representative from Texas called the race a “circular firing squad.” Super PACs and party supporters spent millions on both Gonzales’s and Herrera’s campaigns, making TX-23 one of the most expensive primaries in the U.S. The delayed runoff gave out-of-state interests the time and information needed to fuel an extremely negative dog-eat-dog campaign.

Gonzales won by just 407 votes in the runoff. 16,000 fewer voters participated than in the first round.

“Runoffs are just brutal,” Gonzales remarked after his win. The data seems to agree.

Ranked choice voting solves the problems with runoffs

Primary runoff elections have a worthy goal: ensuring election winners have majority support. However, in practice, they often fail to achieve this, and cost valuable taxpayer dollars in the process. It is clear that change is necessary. At the very least, lawmakers should reduce the amount of time between the initial primary and runoff. However, there is a more powerful reform that improves on the runoff system: ranked choice voting (RCV).

Ranked choice voting captures voters’ preferences without holding two elections. With RCV, voters have the option to rank the candidates in order of preference. If a candidate is ranked first by a majority of voters, that candidate wins — just like any other election. However, if no candidate reaches 50%, an “instant runoff” is triggered. The candidate with the fewest votes is eliminated, and ballots ranking that candidate first will now count for the voter’s next choice. This process repeats until a candidate reaches the 50% threshold.

RCV prevents the turnout decline associated with delayed runoff elections. In an RCV election, the winner always wins more votes in the final round of counting than the first. Compare this to the 81% of runoff winners this year who won fewer votes in the runoff than the initial primary.

With RCV, voters do not weigh in on the final round only if they choose not to rank a top candidate anywhere on their ballot. This impacts 5% of voters in RCV elections, approximately one-thirteenth of the 63% of voters who did not vote in delayed runoffs this year. RCV ensures primaries are decided by a larger, more representative group of party voters, and gives more voters a say between top candidates.

Additionally, by switching to RCV, states can put the money they currently spend on runoffs to better use. The cost of voter education and equipment to run RCV elections is only a fraction of the cost of runoffs.

Research shows that RCV can also facilitate positive campaigning. If nominees are chosen on primary day, the incentive to earn a spot in the runoff by demonizing potential threats — and then to win the runoff by bashing their strongest threat — is mitigated. Moreover, with RCV, candidates benefit from being the backup choice of their opponents’ supporters (as opposed to being the “lesser of two evils” in a runoff).

Alabama, Arkansas, Georgia, Louisiana, Mississippi, and South Carolina already use RCV to help military and overseas voters participate in runoffs. Because of the challenges of mailing runoff ballots overseas so shortly after the primary election, these voters get to rank their choices during the primary. If a runoff election occurs, their vote counts for whichever finalist candidate they ranked highest. All voters in runoff states should be given the same convenience.

Data sources

The data used in this report was obtained from Secretary of State and Election Board websites for election results, through FOIA requests for election costs, and through L2 VoterMapping for voter demographic information.

Our full data set of runoff election results is available here, and our data set of runoff election costs is available here.

About the authors

Ben Fitzgerald was a Research and Policy intern with FairVote. He worked with FairVote through a partnership with Haverford College’s Center for Peace and Global Citizenship (CPGC). He is currently pursuing majors in philosophy and computer science at Haverford College.

Rachel Hutchinson is a Senior Policy Analyst at FairVote. She is an experienced researcher and writer in the electoral reform space. Rachel has previously worked at the Institute for Political Innovation (IPI) and Democrats Abroad. Rachel has a bachelor’s degree in Politics and Sociology from Newcastle University.