The Party That Can’t Say NoHow Fifty Years of Primary Reform Left America’s Parties Without a Way to Vet a Candidate
I wrote a shorter piece earlier this year about the parties’ habit of cultivating “useful idiots,” members who are legislatively compliant but intellectually thin. It did not travel far, and I think I know why. Complaining about weak members of Congress is a personal argument. It names no cause and offers no fix. The real story is structural, and it goes back further than most people realize. This year alone gave us a socialist primary wave inside the Democratic Party, a Senate nominee in Maine who won his primary and then had to quit weeks later, a House member the Republican majority initially voted to protect from expulsion despite a federal ethics finding, and a sitting senator publicly asking his own party to remove his former son in law from Congress. None of that is coincidence. All of it traces back to the same missing step.
Where it started
In 1968, the Democratic Party nominated Hubert Humphrey for president without his winning a single primary. Eugene McCarthy had shown real strength with voters that year, but party insiders chose Humphrey anyway behind closed doors. The convention in Chicago collapsed into chaos, with televised violence outside the hall and open warfare inside it. In February 1969, Democratic National Committee chairman Fred Harris appointed a Commission on Party Structure and Delegate Selection, known afterward as the McGovern Fraser Commission after its chairmen, Senator George McGovern and Representative Donald Fraser. The commission’s own 1970 report, Mandate for Reform, stated its finding plainly: “meaningful participation of Democratic voters in the choice of their presidential nominee was often difficult or costly, sometimes completely illusory, and, in not a few instances, impossible” (Mandate for Reform, Teaching American History).
The commission’s own numbers made the case. In one state, filing fees for a full slate of primary candidates reached $14,000. In one delegation, hospitality fees charged to convention delegates reached $500. Only 13 percent of convention delegates had incomes under $10,000 a year, at a time when 70 percent of the country did. Women held 13 percent of delegate seats, and only one of the convention’s fifty five delegations, which included the states plus the District of Columbia and the territories, had a woman as its chairman (Mandate for Reform). The commission wrote eighteen guidelines meant to open the process, and within a decade the number of states using binding primaries jumped sharply. Delegate selection stopped being a party insider’s job and became a voter’s job.
That was the right instinct applied without a backstop. Party bosses had abused their gatekeeping power in 1968, so the fix removed the gate rather than reforming it. Nobody replaced the vetting function the bosses used to perform, however badly. We got a more democratic process and a less curated one, and fifty years later we are living with the tradeoff.
The spread to open and jungle primaries
The logic of 1972 kept extending. States began experimenting with primary formats that removed party control even further. California’s Top Two Candidates Open Primary Act took effect January 1, 2011. Under it, every candidate for a given office, regardless of party, appears on one ballot, and only the top two vote getters advance to the general election even if both belong to the same party (California Secretary of State). Washington voters passed a similar system, Initiative 872, in 2004 with almost sixty percent support. State parties sued, and the law sat under injunction until the Supreme Court upheld it in March 2008, after which Washington ran its first top two primary that year (Washington Secretary of State). Louisiana runs its own version. Its secretary of state describes it plainly on the state’s own election website: “Most candidates in Louisiana are elected by majority vote in an open primary (jungle primary) system,” where every candidate for an office appears on one ballot regardless of party and only a majority winner avoids a runoff (Louisiana Secretary of State).
Supporters argued that removing party control this way would produce more moderate candidates, since a candidate could no longer just satisfy a partisan base and had to appeal to a broader electorate to finish in the top two. The broader academic record is not so tidy. A widely cited study by Eric McGhee, Seth Masket, Boris Shor, Steve Rogers, and Nolan McCarty examined more than 18,000 state legislators across every state legislature from 1992 through 2008 and found that primary openness has little or no consistent effect on how ideologically extreme the winners turn out to be. Where an effect existed, it often ran the opposite direction from what reformers wanted, producing more liberal Democrats and more conservative Republicans in more open systems (McGhee, Masket, Shor, Rogers, and McCarty). Changing who gets to vote in a primary does not reliably change who wins it. That is an uncomfortable finding for reformers on both sides of this debate, and it points toward a different diagnosis than simply which primary format a state uses.
Splitting the vote, diluting the label
This is where the diagnosis has to shift, even though the primary format studies are mixed. The problem is not just who can vote. It is who can run and what they are allowed to call themselves once they do. A crowded field of candidates who each self identify as a Democrat or a Republican, then attach a modifier such as Democratic Socialist or MAGA, can let an energized, well organized minority take control of a party’s ballot line in a low turnout primary, then carry the full party’s brand into a general election most voters never had a hand in shaping.
The 2026 cycle put this on live display. The Democratic Socialists of America backed 150 candidates this cycle, and by late June more than 30 had won or advanced from their primaries, including upset wins over sitting members of Congress (Washington Examiner). The organization tracks its own results in real time on its own site, listing race by race wins going back to May (DSA Election Night 2026). The most watched of these came on June 30, when 29 year old first time candidate Melat Kiros defeated 15 term incumbent Representative Diana DeGette in Colorado’s 1st Congressional District (NPR). Three DSA backed candidates also won New York congressional primaries that same month, unseating the chairman of the Congressional Hispanic Caucus among others (Washington Examiner). The wave is not universal. On August 4, DSA backed candidates lost several marquee races, including former Representative Cori Bush’s attempt to reclaim her Missouri seat (Newsmax). But the pattern holds often enough to matter. A party primary with low turnout does not require a majority of the party’s own voters to hand over its ballot line. It requires an organized plurality willing to show up, and once that line is handed over, every candidate up and down the ballot inherits whatever label the winner carries into November whether the rest of the electorate agrees with it or not.
This is precisely the failure mode of first past the post voting applied to a multi candidate or low turnout field, and it is the same dynamic that has pushed reformers toward ranked choice voting in some states as a partial fix. But ranked choice only addresses the mechanics of counting votes after the fact. It does nothing to stop a party’s own brand from being captured by a faction with no real intention of representing the median voter that brand was built to attract.
The cost of a real alternative
Our two party system persists partly because building a genuine third option is prohibitively expensive and slow, not because voters are satisfied with only two choices. An independent presidential candidate must gather roughly one million signatures nationwide to appear on all fifty state ballots, a process Robert F. Kennedy Jr.’s own 2024 campaign estimated would cost between fifteen and thirty million dollars (NPR; USA Today). Individual states pile their own thresholds on top. Alabama law requires a petition signed by at least three percent of the qualified electors who voted for governor in the last general election, which for the 2026 cycle works out to 42,458 valid signatures statewide (Alabama Secretary of State). Arizona law sets the bar at one and one third percent of the vote cast for governor in the last gubernatorial election, spread across at least five counties, with at least ten percent of signatures coming from counties under 500,000 people (Arizona Legislature).
None of that is an accident. State ballot access law was written by legislators from the two parties that benefit from keeping the field closed. The practical result is that frustrated voters and ambitious candidates have nowhere to go except into one of two increasingly stretched tents, which is exactly why the big tent theory keeps failing. A tent has to hold together people who agree on very little, and when there is no real exit option, factions fight for control of the tent itself rather than building their own.
Debate stages built for television, not for judgment
The presidential primary debate has become the clearest example of a process optimized for attention rather than for evaluating who could actually govern. In 2015, seventeen major Republican candidates entered the race, and Fox News solved the logistics problem with its own official criteria for the first debate: candidates had to place in the top ten of an average of the five most recent national polls, with a separate undercard debate for everyone else. By the time the field reached South Carolina in February 2016, CBS News set its own qualifying rule for that debate: a candidate needed a top five finish in New Hampshire, a top three finish in Iowa, or a top five average in national and South Carolina polling with at least three percent support somewhere (CBS News criteria, reproduced by the American Presidency Project).
The Democrats did the same thing in reverse scale during their twenty nine candidate field in 2020. For the tenth debate, held in South Carolina, the Democratic National Committee’s own published criteria required a candidate to either win a pledged delegate in Iowa, New Hampshire, or Nevada, or clear ten percent support in four separate polls, or twelve percent support in two South Carolina polls from a list of thirteen approved pollsters, all inside a three week qualifying window (Democratic National Committee).
A field that large rewards whichever candidate finds the sharpest, most inflammatory way to stand out in a ninety second answer. It does not reward the candidate best equipped to run the executive branch. The Brookings Institution has made this point directly: current debate qualification rules measure a candidate’s donor count and polling support, not the candidate’s capacity to govern. Nearly every licensed profession in this country, medicine, law, even a manicurist, requires some form of peer evaluation before practice. American politics asks for none (Brookings).
Opposition research is not vetting
There is a point in this argument that deserves more attention than it usually gets. Vetting a candidate today mostly means waiting for the other side’s opposition research shop, or a reporter, to dig up something damaging after the fact. That is not vetting. It is discovery, and it happens on somebody else’s timetable and for somebody else’s benefit, not the voter’s. Three cases from the past three years show exactly what that failure looks like in practice.
The academic research on negative advertising explains why this substitute does not work. A 2007 meta analysis by Richard Lau, Lee Sigelman, and Ivy Brown Rovner, published in the Journal of Politics, reexamined the entire research literature and concluded that “the research literature does not bear out the idea that negative campaigning is an effective means of winning votes, even though it tends to be more memorable and stimulate knowledge about the campaign,” while it does “slightly lower feelings of political efficacy, trust in government, and possibly overall public mood” (Lau, Sigelman, and Rovner, The Journal of Politics). A separate study by John Sides, Keena Lipsitz, and Matthew Grossman, published in American Politics Research, found that voters can recognize an ad as negative without believing it teaches them anything useful, and the two judgments are unrelated in most voters’ minds (Sides, Lipsitz, and Grossman, American Politics Research). Negative advertising generates noise and depresses trust. It was never designed to answer the one question that actually matters, which is whether a candidate is fit to hold the office and will represent the party’s platform once there.
Maine’s 2026 Senate race shows what happens when nobody answers that question before the ballot is printed. Graham Platner, an oyster farmer and combat veteran with no prior political experience, became the frontrunner after Governor Janet Mills suspended her own campaign in late April while remaining on the ballot. Reporters had already surfaced a tattoo on Platner’s chest resembling a Nazi symbol and a history of deleted Reddit posts before voters went to the polls. None of it stopped him. He won the Democratic nomination on June 9, 2026, with roughly 72 percent of the vote to Mills’s 19 percent (Bloomberg; The New York Times). A month later, a woman named Racicot came forward with an allegation that Platner had raped her. Senator Bernie Sanders, Senator Elizabeth Warren, Senator Ruben Gallego, and Representative Ro Khanna all withdrew their endorsements within days (Reuters). Platner announced he was suspending his campaign on July 8 and formally filed his withdrawal with Maine’s Secretary of State on July 11, clearing the way for the state Democratic Party to select a replacement through a nominating convention by July 27 (NPR; NBC News). The warning signs existed before the primary. The party had no mechanism to act on them until after he had already become its nominee in a race that could decide control of the Senate.
Congress has its own version of the same failure, and it shows what happens once a flawed nominee is already sitting in office and a party has to decide whether removing him is worth the cost. Representative George Santos of New York was elected in 2022 on a resume later shown to be substantially fabricated. On November 1, 2023, the House voted on a resolution to expel him before the Ethics Committee had even released its final report. It failed, 179 to 213, with House Republicans voting 24 in favor and 182 against (U.S. House Clerk, Roll Call 564). Two weeks later, the House Ethics Committee released a report finding substantial evidence that Santos had broken federal law, including converting campaign funds to personal use, as reported at the time by the network carrying the full text (PBS NewsHour). Only after that report, and after weeks of mounting public pressure, did the House vote again, on December 1, 2023, and this time expel him 311 to 114, making him the sixth member in the chamber’s history removed by his colleagues (BBC). For a full month, a chamber controlled by a narrow Republican majority chose to keep a member accused of serious wrongdoing in his seat rather than give up a vote, and only public pressure following an official government finding changed that calculation.
The most personal version of this failure is unfolding in Ohio right now. Senator Bernie Moreno, a Republican, spent August publicly calling on his own party to remove Representative Max Miller, also a Republican and Moreno’s former son in law, from Congress. The allegations came from Moreno’s daughter, Emily, Miller’s ex wife, who said in court filings that Miller had pointed a gun at her head, poured hot water on her, and caused an injury to their infant daughter (The Guardian). Moreno called Miller “a danger” and said he “fails” the basic standards required to hold office (PBS NewsHour). On August 4, 2026, the House Committee on Ethics confirmed in its own official statement that it was reviewing allegations Miller “may have engaged in domestic violence and abuse” and “illegal drug use” under its own review rule (Committee on Ethics). Miller has refused to withdraw from his reelection race. If a sitting senator with direct family knowledge of a candidate’s conduct has no formal way to flag it before that candidate is renominated, and can only make his case in public after the fact, then no meaningful vetting exists inside either party even when the concern comes from inside the family.
What other systems do differently
Parliamentary democracies solve this problem with a step that American parties gave up decades ago, and it is worth looking at closely rather than dismissing as un American. In the United Kingdom, the Conservative, Labour, and Liberal Democrat parties all run centralized, professionalized assessment processes before anyone can stand for Parliament under the party’s name. The Conservative Party’s Parliamentary Assessment Board, for instance, is an assessment center that evaluates prospective candidates on standardized, competency based criteria before they even reach a local selection meeting. Research on the process suggests it operates in a genuinely objective way, since the share of women passing the board roughly matches the share applying (Institute for Government). Voters still choose the winner. The party simply retains the right to decide who is allowed to compete for its label in the first place, which is a form of vetting no American primary performs.
Parliamentary systems also give minority factions a path that does not require them to capture or fracture a larger party. A faction can form its own party, win a modest number of seats, and then join a governing coalition, trading policy influence for legislative support without ever pretending to be something it is not on the ballot. American politics has a domestic cousin to this idea that is almost never discussed: fusion voting. New York’s own election law allows a candidate to be nominated by more than one party or independent body at once, and if that candidate is not an enrolled member of a party that nominates him, he must formally accept the separate nomination for it to count (New York State Senate). Under that same law, the candidate’s name can appear on the general election ballot once for each party line that nominated him, and the votes cast for him on every line are added together toward one total (New York State Board of Elections). A minor party can cross nominate a major party’s candidate, giving its own voters a way to support that candidate without pretending to be Democrats or Republicans, while keeping its own vote total and identity fully separate.
There is also a lighter version already used in Congress. Senators Bernie Sanders of Vermont and Angus King of Maine are both registered independents who caucus with the Democratic Party for committee assignments and organizational votes, without ever adopting the Democratic label on a ballot. This is functionally identical to a parliamentary coalition partner. The party gains a working majority vote without diluting its own brand by handing its ballot line to someone who never joined it, and the officeholder keeps an independent identity that his voters actually chose.
Where this leaves the argument
So has the democratization of party politics gone too far. The evidence suggests the honest answer is not that voters were given too much say. The McGhee, Masket, Shor, Rogers, and McCarty study is fairly persuasive that opening up who can vote in a primary does not reliably produce more extreme or more moderate winners on its own (Princeton). The real failure sits one step earlier in the process, in the disappearance of any institutional check on who is allowed to compete for a party’s name and resources before the voting even starts. Platner cleared a primary with warning signs already public. Santos kept his seat for a month after a resolution to remove him failed on a party line vote, before an ethics finding forced the party’s hand. Miller remains on the ballot while his own former father in law, a sitting senator, says publicly that he should not hold office. A socialist primary wave is reshaping the Democratic ballot line in dozens of districts without any party structure asking whether the winners represent where the broader electorate actually sits. We took away the boss’s veto in 1968 for good reason, and we never built anything to replace its one legitimate function, which was making sure a nominee had actually earned some credibility with people who knew the office and the platform.
Four changes worth making
None of these require rolling back the primary vote, and none of them hand power back to smoke filled rooms.
First, parties should build a real, transparent candidate certification process before granting anyone the right to run under their name, modeled loosely on the assessment boards British parties already use. This would not block anyone from running. It would simply mean that using the “D” or “R” line on a ballot requires clearing a minimum bar set by people who know what the job actually requires, rather than being automatic for anyone who files paperwork and pays a fee. A basic background check for the kind of conduct that surfaced in Platner’s and Miller’s cases would be a start, not an overreach.
Second, expand fusion voting beyond New York and the handful of other states that currently allow it. This gives ideological factions, whether a Democratic Socialist wing or a populist wing, a real ballot line of their own to build support around, without needing to capture or dilute the larger party’s nomination to have any voice at all.
Third, formalize caucusing rules the way the Senate already treats Sanders and King, so that state legislatures and even future presidential tickets have a clear, workable path for coalition partners who are not full party members but still want to govern alongside one. This is the American version of a parliamentary coalition, and it already works.
Fourth, reform debate qualification standards so that a portion of stage access depends on some form of peer or professional evaluation, not purely on donor counts and polling. The goal is not to gatekeep ideas out of the race. It is to stop rewarding the single most inflammatory answer in a ninety second window as the only reliable way to get noticed in a field of fifteen or more candidates.
Four changes worth making follow from this record: a real certification process before anyone gets the right to run under a party’s name, wider use of fusion voting so a faction can build its own ballot line instead of capturing someone else’s, formal caucus rules like the ones Sanders and King already use, and debate standards that reward some form of professional judgment rather than donor counts and polling alone. None of this rolls back a single voter’s right to choose. It restores a step that used to exist before 1968 and that nothing has replaced since: someone inside the party actually vouching that the candidate belongs there.
