The 2014 Primaries: Why They Matter, What’s Wrong With Them, and How to Fix Them

Released May 20, 2014

The 2014 primary season is well underway. Seven states and the District of Columbia have already held their primaries, with six more voting today, and many more will be holding primaries over the next few weeks. Primaries typically get little coverage, but they are one of the most important – and problematic – aspects of the American electoral system. For information on how primaries work in your state, see FairVote’s overview of primary laws and rules on who can participate in which primaries in the 50 states.


Notable Facts: 

  • Primary election turnout has plunged in the United States. According to a 2010 report by the Committee for the Study of the American Electorate, the average turnout rate of the voting age population in the three congressional midterm elections held from 1962 to 1970 was more than 30%. In contrast, the average turnout in the primaries in the three midterm elections held from 2002 to 2010 was only 17.7%.
  • Most general elections are “no contests.” Using our unique projection model, which was 100% accurate in its 2012 projections, FairVote can confidently project outcomes in general elections in 371 of 435 congressional districts in 2014, months in advance of the election. In the vast majority of cases, the primary election is the only election where change is possible.
  • Safe seat winners can earn office with microscopic vote totals. In 2012, Marc Veasey won the Democratic primary runoff in the safely Democratic 33rd congressional district in Texas with the votes of less than 3% of the voting-eligible population.


What are primaries anyway?

First adopted in the United States less than a century ago, primary elections are the method by which states allow parties to nominate candidates to appear on the general election ballot for most elected offices. Each state has different rules for how primaries are structured and who gets to participate in them. The most common system is one in which both major parties (Republican and Democratic) hold elections on the same day a few months before the general election, and the candidate with the most votes is nominated by the party and advances to the next election.

The United States is alone in conducting internal elections for political parties. Some major parties in other nations, such as France, are starting to explore more participatory ways of selecting their nominees, but the parties administer and pay for these selection processes themselves.

Why do I need to care about primaries?

In theory, primaries are just the first round of the election. Most people don’t care about them because they assume the decisive election – the one that determines who will be elected to office – is in November. But in reality, the primary election is decisive most of the time. That’s because most districts and states favor one of the two major parties so strongly that whichever candidate wins the favored party’s primary will be certain to win in November. For instance, FairVote’s Monopoly Politics 2014 report shows that 371 of the 435 congressional districts will be safe for the district’s majority party this November, and that 326 districts (three out of every four House races) would be safe for the majority party even without an incumbent.

So primaries are often decisive. What’s wrong with that?

 FairVote believes that voters should always have the ability to influence the outcomes of general elections. While reforms that focus solely on primaries can be important to making our election process more inclusive and reflective of voter preferences, they are not a substitute for making sure that the decisive election is in November, when voter turnout is highest and most representative.

Decisive primaries would be less problematic if voters were willing (and able) to participate in them. But turnout in primaries is consistently low. In 2012, the Bipartisan Policy Center reported that, even with the inflation of presidential primaries, average primary turnout was 15.9% of the voting-eligible population in the 45 states that held statewide primaries in 2012. Turnout in the general election, on the other hand, was 58.2% of eligible voters – not fantastic by any means, but a whole lot better than in the primaries.

But even that disparity understates the severity of the problem. Since only one primary contest really matters in most districts and states, the turnout in the decisive primary is usually less than 10% of eligible voters in that jurisdiction. Average turnout in Democratic primaries was 7.5% of the voting-age population in 2012, for example. So in a safely Democratic district with a competitive primary in which the winning candidate receives about half the vote, candidates can be elected to Congress with the votes of less than 5% of eligible voters. See the example of Rep. Veasey above, who needed less than 3% of the voting eligible population to secure his seat.

There is no indication that this trend will reverse in this year’s primaries. In Texas, for instance, turnout in the March primary was just 11.7% of all eligible voters – and, of course, even lower participation for each individual congressional party primary in decisive safe seat elections.  

Worse, the primary system has distorting effects on the makeup of Congress. Evidence, including this detailed analysis of primary electorates in Bay Area cities in California, suggest that the primary electorate is far less representative of Americans than the general election electorate; essentially, it is older, whiter, and wealthier. Candidates who advance to the general election are more likely to represent the ideological middle of the most active members of their party, rather than the middle of the district as a whole.

When plurality rules are used in the primary (that is, when the candidate with the most votes wins even if he or she does not have majority support), extreme candidates can be elected when moderate candidates split the vote. For example, controversial 2012 Senate candidate Todd Akin won the Republican nomination with just 36% of the vote, defeating two rivals who may have split the majority vote in part due to Democratic tactics designed to help Akin win. Akin’s congressional career began with an even lower plurality win in his first House primary in 2000, when he earned just 26% of the vote in a safely Republican seat.

The distorted incentives of the primary system can be held directly responsible for the tendency of Members of Congress to play to the extremes of their bases, as in the government shutdown of October 2013.

That does sound pretty bad. What has been done to address this?

A variety of solutions have been tried to fix the primaries. None have worked out that well. Here are a few:

Open primaries: Many states use open primaries instead of closed primaries – that is, party primaries in which people can vote even if they are not registered with the party conducting the primary. The hope is that party primaries that are opened to the full electorate of a district will elect more moderate candidates.

Open primaries are better for independent voters, since they guarantee that they will have an opportunity to participate in the decisive election regardless of their political affiliation. However, political science research has shown that open primaries do little to elect more moderate candidates in part because primary contests are of greatest interest to highly partisan voters and because most major party voters will stick with their own party’s race even if their nominee is destined to lose in the general election.

Primary runoffs: A handful of states use runoffs in their primary elections, holding a second election a few weeks or even months after the first if no candidate receives above a certain threshold of the vote (usually 50%) in the first round. Runoffs might solve the problem of extreme candidates being elected with small shares of the vote, but they also exacerbate the already severe turnout problem in primaries. As a 2013 FairVote report showed, turnout decreased 35% on average between primaries and primary runoffs held from 1994 to 2012.

Top Two: California, Washington, and Louisiana all use a version of the Top Two primary system. In Top Two, all candidates of all parties appear together on the primary ballot, and the two candidates who receive the most votes in the primary advance to the general election (in Louisiana, the “jungle primary” actually takes place on Election Day and the runoff in December). Advocates of Top Two argue that it will help more moderate and representative candidates be elected, because voters in the high-turnout general election will sometimes get to choose between two candidates of the same party rather than facing the usual choice of a Democrat and a Republican. In theory, the more moderate candidate would usually win this election by receiving votes from the party without a candidate on the ballot.

But political scientists have found that Top Two does not have a positive effect on moderation either. It has little impact on who advances to the general election, and – as FairVote’s reports on the Top Two systems in place in Washington and California have shown – general elections almost always feature one Democrat and one Republican. In Washington State, for example, 27 of 28 Top Two general elections for the House have featured one Democrat and one Republican. Yet when two candidates of the same party are featured, many voters from the party that are not represented on the ballot will skip that race.

Top Two almost always excludes third party and independent candidates from the general election ballot Most disturbingly, vote-splitting in the primary round sometimes leads to a decisive election in which the party that most people in the district prefer is not on the ballot – as was the case in California’s majority-Democratic 31st district in 2012, where voters could choose only between two Republicans on Election Day.

Does FairVote have any better ideas?

Yes! One better idea is a variant on Top Two is the Top Four primary that a Colorado reformer is now trying to place on the November ballot. Top Four keeps the “general primary” of Top Two, where all candidates are on the same primary ballot, but instead of just two candidates advancing to the general election, four advance – ensuring that most voters will find someone to their liking on the ballot in November. Then, in the general election, ranked choice voting is used to determine the winner.

In FairVote’s retrospective projections for Top Four in Washington State, for instance, there would still have been 28 of 28 races with both major parties on the ballot in the general election. But there would also have been 25 elections with two candidates from the same party on the ballot, and ten general election races with a minor party or independent candidate.

Ranked choice voting is an improvement with or without an open primary system. It allows voters to vote their preference without hurting their preferred candidate due to vote splitting, and guarantees that the winner will have majority support of all voters who expressed a preference between the two candidates in the last round of the “instant runoff.” The general election also is less likely to resort to negative attacks and name-calling, as suggested by our ongoing study on the impact of RCV on the civility and tone of campaigns.

That doesn’t totally solve the problem of uncompetitive districts where one party wins all the time, does it?

Top Four would help, but it’s not the ultimate solution. It would make more districts and states with one-party control competitive in general elections, though not when the majority party rallies around a single candidate. It also fails to address the issue of diversity in representation – that is, making sure our democracy is truly representative of the people. 

To provide “open” general elections, we need to get to the root of the problem: winner-take-all rules. Under winner-take-all, 51% of voters in a district can win all of that district’s representation. In today’s highly partisan climate, most elections with one winner will inevitably be safe for one party in the general election. For example, there are 73 districts in the 10 states that one might say are “non-gerrymandered” because that state has only one district or because the state is one of the three (Arizona, California and Iowa) where an independent commission drew lines with limited or no use of political data. Despite these “non-gerrymandered lines,” we project that 58 of their 73 districts will be safe for one party in November – that’s nearly four in five districts.

What we really need is three basic electoral changes: 1) fewer but larger districts; 2) electing more than one candidate in these districts; 3) using a fair representation voting system. In fair representation voting, like-minded voters elect candidates in proportion to their voting strength. In a five seat district, 51% of voters would elect three seats, not all five, while 49% of voters would elect two seats rather than none. Under this system, a group of like-minded voters making up 17% of the district could elect a candidate of choice.

To learn more about how this system would work, see our districting plan for the whole country and examples of fair representation systems already used in our local elections at FairVote’s plan could be implemented by federal and state statute without a constitutional amendment. It would give every voter in every state with at least two seats a competitive election, and it would break up one-party representation in every multi-seat district with at least three seats. 

Fair representation voting wouldn’t just affect general elections, though – it could also be used in party primaries. In fair representation primaries, parties would nominate multiple candidates in each multi-member district. Parties might choose to nominate five candidates to run in a five-seat district, for example. Those five candidates would be determined using a fair representation system, meaning that each would need to receive about 17% of the vote to be nominated. Using such a system would allow a much broader diversity of candidates to appear on general election ballots, as candidates from all wings of the “big tent” parties could be nominated rather than just the generic party candidates that are typically selected today. More women would likely run and win and far more people of color would elect preferred candidates.

As detailed in Monopoly Politics 2014 and the Fair Voting Solution, fair representation voting would have a major impact on the ideological makeup of Congress. Instead of a bias toward Republicans and a vanishing center, fair representation voting would produce balance between the two major parties and a robust cohort of “moderates” – including representatives who fall outside of the traditional left-right spectrum.

Fair representation voting would make both primary and general elections more representative and competitive. Without it, we will never address the deepest structural flaws in the American primary system.

Check back at over the course of the summer for more analysis on the 2014 primaries