Voters need help: How party insiders can make presidential primaries safer, fairer, and more democratic

Raymond J. La Raja and
Ray La Raja
Raymond J. La Raja Professor, Political Science - University of Massachusetts-Amherst
Jonathan Rauch

January 31, 2020

Presidential-nominating contests in both major political parties are at risk of producing nominees who aren’t competent to govern and/or don’t represent a majority of the party’s voters. Raymond La Raja and Jonathan Rauch argue this is a result of the declining role of party insiders in the nomination process and call for the reversal of that trend. Primaries function best, they claim, when voters and party professionals work in partnership.

Table of Contents

I. Introduction
II. The history of presidential primaries
III. Problems with the present process
IV. The promise of professional vetting
V. Proposed reforms

Campaign signs are seen outside a polling station on the last day of early voting in Dallas, Texas, U.S., November 2, 2018. REUTERS/Mike Segar

I. Introduction

A game of chance

“He is the chaos candidate, and he would be a chaos president.” Jeb Bush’s famous and amply fulfilled warning from December 2015, aimed at Donald Trump, might just as accurately have included a further phrase: “And our nominating system is a chaos process.”

In the Republican presidential nominating cycle four years ago, 17 candidates, ranging from governors and senators to a cardiac surgeon, a tech executive, and, of course, a reality TV star, entered the fray. The TV celebrity won, despite never receiving a majority of the Republican primary vote until his lead was insurmountable, because the field was too fragmented to coalesce around an alternative. Might something similar happen in the Democratic Party’s nominating process this year? Might the process devolve into chaos and factional capture? The answer is yes.

That is not a prediction. As was also true for the Republicans four years ago, odds are on the side of a less chaotic outcome. Importantly, however, nothing in the process guarantees one. The same forces which hijacked and disrupted the Republican process are hard at work on the Democratic side—now and in the future. Both parties’ presidential-nominating contests have reached a point where they cannot promise to choose nominees who are competent to govern or who represent a majority of the party’s voters.

In this essay, we argue that those flaws are inherent results of party insiders’ demotion to spectator status. Reforms to increase voter participation or reduce the influence of money (for example) have pros and cons, but they can no more address the absence of professional input than putting more passengers on airplanes or adding flights can remedy a shortage of air-traffic controllers. Process reforms are not substitutes but complements for professional judgment and organizational skill. Primaries function best when primary voters and party professionals work in partnership; two filters are better than one, and in fact, neither filter works well by itself.

More specifically, we argue that:

  • Professional input makes the process more representative. Copious theory and evidence, dating back to the time of America’s Founders, show that nomination by plebiscite (popular vote) can collapse into randomness or minority capture, and it does not dependably aggregate and reflect the preferences of Democrats and Republicans. When many candidates are in the field, professionals help majorities and coalitions to form, and they help prevent minorities and factions from capturing the process.
  • Professional input strengthens quality control. Primary elections place insufficient emphasis on evaluating nominees with an eye toward competence at governing: that is, selecting individuals with traits such as coalition-building skill, connections to varied constituencies, ability to work with others, and IOUs to and from other politicians. Only professionals can fill that gap.
  • Professional input deters renegades. Combined with the party ballot’s accessibility to all comers, the plebiscitary nomination process opens the field to demagogues and charlatans. Party leaders have strong incentives to keep candidates off the party ballot who are dangerous to both the party and democracy.
  • Professional input checks the power of billionaires and media elites. Influence in nominations has shifted dramatically toward actors who bear no responsibility for governing. Billionaires can bankroll themselves or favored candidates, while media elites propel those who break norms and generate conflict. Party professionals tend to favor candidates who are responsible to broad voter constituencies and other members of the governing party.
  • Professional input is widely acceptable to Americans. There is nothing undemocratic or un-American about professional vetting of nominees. To the contrary, even as the nation became more democratic and inclusive since its founding, formal and informal vetting of candidates by parties and professionals remained standard practice until just a decade or so ago. Even today, survey results indicate that most Americans support giving parties and professionals a voice in the process.
  • Restoring professional input is mechanically easy, but politically hard. Methods might include superdelegates, early votes of confidence, ratings or signoffs by party stalwarts, use of influence on debate participation, unbinding convention delegates, routing more campaign money through the party organizations, and many more possibilities. The harder challenge is pushing back against democracy fundamentalism, the idea that more democratization is always good for democracy—something which the Founders knew is not true.

We do not claim that primary elections have no place or serve no purpose. To the contrary: They test candidates’ abilities to excite voters, raise money, and campaign effectively; they provide points of entry for fresh faces, ideas, and constituencies; they force candidates to refine their messages and prove their stamina. What we do claim is that primaries are insufficient. By themselves, they are only half of a functional nominating system. Without professional input, the nominating process is fraught with risks that turn filling the country’s highest office into game of chance. If Democrats don’t think it can happen to them, they are deluding themselves.



Picking presidential candidates has never been pretty. The nomination process was not designed rationally over its 200-year history. Instead, it is a jumbled contrivance, pieced together by ambitious candidates, partisan factions, and self-interested politicians. In the early years of the Republic, members of Congress nominated candidates for the presidency, precisely to stymie factional candidates. By the election of 1828, however, the egalitarian spirit of the Jacksonian Age pushed the selection process into state and national conventions populated by local party professionals. By the turn of the 20th century, Progressives saw conventions themselves as instruments of “boss rule.” They pushed for primaries for all levels of office, even though few states chose to use them for presidential nominations. Nonetheless, the mixed system of choosing delegates through party primaries and insider-controlled caucuses provided some balance, with party leaders using primaries to assess the appeal of plausible candidates before choosing the nominee at the convention.

The next significant populist push came in the aftermath of the tumultuous 1968 Democratic convention, where Hubert Humphrey was coronated despite not having competed in any of the primaries (though surrogate candidates, so-called favorite sons, did compete on his behalf). In the aftermath, Democratic leaders acknowledged the party’s legitimacy problem. In 1969, the Democratic National Committee (DNC) organized the McGovern-Fraser Commission, which changed the rules to favor primaries. Republicans followed the Democrats’ example, without much debating the implications.

“[N]o other major democracy routinely uses primaries to select party nominees.”

Using plebiscites to choose party nominees was unprecedented among the world’s democratic political parties. Indeed, no other major democracy routinely uses primaries to select party nominees. Without much forethought, the reforms of the 1970s weakened party elders’ capacity to pick a prominent, unifying, or broadly representative national figure to lead their ticket. Byron Shafer’s seminal research revealed how this “quiet revolution” (his book’s title) transformed power relationships within the parties and political leadership itself.

The link between rules and outcomes soon became evident. The new system produced a weak candidate in 1972 (George McGovern) and a weak president in 1976 (Jimmy Carter), and it generated strife over delegate rules in the 1980 election when Sen. Ted Kennedy challenged President Carter. In 1981, the DNC’s Hunt Commission sought to improve the vetting process by creating a set of delegates who had experience in politics (many were office holders) and who would not be pledged automatically to candidates based on primary outcomes—a rare instance of retrenchment from populism. These so-called “superdelegates” included members of Congress, governors, and Democratic party officials: political careerists with skin in the game.

Superdelegates served conscientiously and never came close to countermanding primary voters. Much of their influence was at the front end of the process, with candidates having to take their views seriously as a key constituency representing both voters and the DNC. Nonetheless, in the aftermath of the 2016 election, populism prevailed again. This time supporters of Sen. Bernie Sanders protested what they believed had been Hillary Clinton’s unfair advantage in securing superdelegates’ support (even though it was the voters, not the superdelegates, who chose Clinton). In a contentious vote, DNC members adopted new rules barring superdelegates from voting on the first ballot at the nominating convention, which ordinarily means they will not vote at all. The weakening of superdelegates’ role reduced, yet again, experienced professionals’ opportunity to weigh in on their party’s nomination.

“As recently as 2008, leading political scientists argued that decision-making remained in the hands of party grandees. But then came 2016.”

Until recently, democratization appeared cost-free: all upside, no downside. In the 1980s and after, the plebiscitary system seemed to nominate candidates of high quality. In fact, both parties’ nominees might also have won under the pre-reform “boss” system. They were mostly experienced politicians who had deep connections to other professionals. Democratic nominees after 1980 included two vice presidents, two successful governors, and three prominent senators (albeit one with little national experience). On the Republican side were a vice president, two successful governors, and two prominent senators. All were acceptable to their party establishments and to broad cross-sections of the party’s base.

What often went unnoticed, however, was why the nominees were so solid: Professional input had survived informally, in the form of a vetting process known as the invisible primary. Even after the rules changed in the early 1970s, candidates still needed to prove their viability, which meant showing they could win influential endorsements, command media attention, appeal to multiple constituencies, and raise money. To be competitive, they had to run a gauntlet of party bigwigs, factional leaders, money brokers, and media gatekeepers. As recently as 2008, leading political scientists argued that decision-making remained in the hands of party grandees.

But then came 2016.



We need not recount here the devastating effectiveness with which Donald Trump’s insurgent candidacy steamrolled the traditional gatekeepers, commandeered media attention, and mobilized what some of his backers called his “troll army.” However, the weakening of gatekeeping was not limited to one candidate or one party. The Democratic Party establishment found itself barely able to contain the insurgency of Sanders, even though he was not a Democrat and he did not win a majority of self-identified Democrats except in his home state of Vermont and neighboring New Hampshire.

Neither candidate changed the system all by himself. Rather, both saw and exploited the invisible primary’s fragility. Candidates could bypass traditional moneymen by reaping donations online, tapping deep-pocketed tycoons, or funding themselves. They could bypass traditional media by using social platforms like Twitter and Facebook, and they could hijack traditional media by behaving outrageously. They could treat their lack of endorsements as a mark of authenticity.

Left to their own devices and the tender mercies of candidate ads and outside groups, voters do the best they can, but their abilities are limited. Voters are often ambivalent and cross-pressured, and their knowledge and bandwidth are necessarily finite. When offered several options, individuals often lack enough information to make a choice that reflects their own preference. In fact, presidential primary voters often vote mistakenly for candidates who do not reflect their views; one important study of the 2008 presidential primaries found that voters barely do better than chance at picking the candidate who best agrees with them. Scholarship points to voters’ heavy reliance on heuristics (simple rules of thumb), which are linked to social identities rather than rational decisions. Summarizing abundant research on voter decisions, the political scientist Bruce Cain concludes that “the original sin of citizenship is our cognitive fallibility; namely, limitations in knowledge and motivation.” In short, the nomination process makes unrealistic demands on voters, not because voters are lazy or stupid—they’re not—but because they are human.

“[T]he nomination process makes unrealistic demands on voters, not because voters are lazy or stupid—they’re not—but because they are human.”

But even if all voters were cognitive virtuosos and informed to the hilt, aggregating individual preferences is inherently much more difficult than people assume. As long ago as the 1780s, the French mathematician and philosopher Nicolas de Condorcet showed that in any field offering multiple candidates, majorities may prefer Smith over Jones and Jones over Brown, yet Brown may nonetheless beat Smith. The reason is that different pairings of candidates produce different majorities, and no majority is the majority. In the 1950s, the Nobel Prize-winning economist Kenneth Arrow took the argument further, proving mathematically that, no matter what system of voting is used or how rationally voters behave, the electorate can fail to arrive at any consistent majority choice, even when choosing from as few as three candidates.

In other words, theorists have understood for many years that there is no one “right” way to aggregate group preferences, nor is there any one representative majority. The candidate who emerges in a multi-candidate race might be picked randomly from a temporary majority simply because of a fluke of election timing or an unforeseen event. In fact, there is a nontrivial likelihood that the plurality winner will turn out to be unwanted by a majority. When the number of candidates reaches double digits or more, elections can enter a world that we think of as Arrow’s nightmare, where the process is not particularly representative of anything.

As if aggregating preferences across many candidates were not challenging enough, the U.S. primary system compounds Arrow’s nightmare by adding its own elements of randomness. For example, the frontloading of primaries in the Democratic 2020 race might allow a candidate to wrap up the nomination in the early states within as little as about five weeks, based on a small and not especially representative share of the electorate. Moreover, because the ballot registers only each voter’s first choice, it provides no additional information about the intensity of preferences for alternative candidates. If majorities split their vote among several candidates competing in the same “lane,” the first choice of a plurality of voters can win even if she is the last choice of a majority. Candidates with widespread approval can get locked out—especially in the winner-take-all GOP—because there is no easy mechanism to translate their broad support into the final selection of the nominee. Under the convention system, by contrast, party leaders could consider the desires of multiple constituencies and move to a more broadly representative second-choice candidate if the plurality candidate was unacceptable to the larger coalition. Party leaders did so on several occasions.

Even under near-optimal conditions, the vagaries we just described give dangerously good odds to fringe candidates. Consider the incentives facing candidates in a large field. To endure the early battles requires finding a lane and mobilizing a loyal faction rather than assembling a broad coalition. Coalition-building necessarily involves making compromises that upset factional loyalists. Think of a Republican candidate vying for the rural vote who favors some mild gun-control measures to attract suburban voters. Her relative moderation might sink her among many rural voters who are strongly pro-gun. When a lot of candidates are in the race, it is often safest to woo purists—who are over-represented as primary voters—by making impractical promises or extreme appeals.

Outsiders, extremists, and demagogues are good at pursuing factions, because they are untethered to the realities of governing. Compromise and coalition-building are logical necessities for crafting and passing legislation. But committing to logical necessities is not always a helpful strategy for winning the nomination under the current rules. The goal is not to win a majority but to survive and hope that luck pits you against a gaggle of candidates who compete with each other in a different factional lane.

“The media can be a powerful accomplice to fringe candidates who play their cards well.”

The media can be a powerful accomplice to fringe candidates who play their cards well. Media coverage heightens the importance of early states because of the narrative drama of uncertainty amid a large field of candidates. That means candidates who happen to win a plurality, or who come close, are catapulted to national attention—never mind that their success came among a small number of not particularly representative voters in a handful of small, not particularly representative states. And extremism, outrage, and conflict are catnip for journalists, and thus for candidates who hope to attract media notice.

In this system, the candidate who emerges may well be one who is intensely disliked by a large group of party voters. Even when choices are few, voting is a blunt instrument for making collective decisions. When choices proliferate, the primary electorate is practically flying blind.

We do not predict that the Democratic nomination cycle in 2020 will replay the 2016 Republican race. Democratic partisans’ desperation to oust the president will encourage sobriety. Nonetheless, Democrats should not feel immune to a Republican-style car wreck. Many of us have had the experience of easing our foot off the gas pedal after passing a roadside accident, only to accelerate miles down the road. Make no mistake. Without better guardrails, the selection process on the Democratic side is unsafe at any speed.


Americans have good reason not to like the current nominating system—and indeed, they do not like it. In a 2016 Pew Research Center poll, just 35% of voters said primaries are a good way of selecting the best-qualified nominees. This view has not changed much since 1992. Not surprisingly, all candidates’ supporters, except Trump’s, viewed the primary system negatively in 2016. But even one-in-three Trump supporters said primaries have not been a good way of choosing nominees. Studies also show that large proportions of Americans favor reforming the presidential nominating process, particularly respondents in states that come in later stages of the selection process.

Residents of late-voting states are understandably concerned about fairness. Early states get most of the attention from politicians and have the biggest impact on the outcome. The sequencing of states—with Iowa and New Hampshire arbitrarily leading off—makes the primary system unrepresentative of rank-and-file voters in a diverse polity. Candidates drop out long before voters in many states can participate. Then there are additional discrepancies in participation in states that use caucuses instead of primaries to select delegates. The caucus method is even more unrepresentative because it requires significant effort to participate, which caters to older voters and those for whom politics is a passion.

“One misconception about reforms that democratize politics is that they increase fairness by taking elites’ finger off the scales. Changing the rules, however, may simply tip the scales toward different kinds of elites.”

One misconception about reforms that democratize politics is that they increase fairness by taking elites’ finger off the scales. Changing the rules, however, may simply tip the scales toward different kinds of elites. More than a century ago, skeptics pointed this out to reformers who claimed that primaries would thwart party bosses. As early as 1909, Henry Jones Ford, a political scholar of the early 20th century, warned that the direct primary would “take advantage and opportunity from one set of politicians and confer them upon another set,” with the result that “its pretense of giving power to the people is a mockery. The reality is that it scrambles power among faction chiefs and their bands.” His conclusion was withering: The direct primary would make politics “still more confused, irresponsible and costly.”

When primaries became widespread and decisive in 1972, a new set of politicians increased their odds of winning significantly: movement candidates and outsiders, including George McGovern, Jimmy Carter, and, of course, Donald Trump. Candidates now worry less about demonstrating to their peers that they have aptitude for governing and more about producing media storylines for journalists who cover the campaign. In 1983, the political scientist Nelson Polsby predicted that the new nomination process would attract candidates with a different set of talents and qualities: “Without the potentiality—or perhaps the threat—of peer review, the incentives for prospective candidates to maximize name recognition and minimize public service—to cultivate the show horse rather than the work horse style—would become overwhelming.” Polsby proved prescient. A plebiscitary system increases the odds of installing ineffective or irresponsible leaders in positions of immense responsibility, while at the same time reducing the ability of party establishments to influence and constrain those leaders’ behavior.

That might be a worthwhile trade if the direct election of nominees improved representation, but it does not. In 2012, Dennis C. Spies and André Kaiser looked at data for 53 parties in nine western European countries between 1970 and 1990. Their finding: “Parties in which party elites decide the nomination of candidates show slightly higher degrees of representation than parties with more inclusive selectorates.” (Italics added.) The reason: Party professionals are more attuned to the need to build coalitions, win general elections, and govern successfully than are the self-selected activists and political hobbyists who dominate participatory selection processes. Research by Byron Shafer and Regina Wagner considers how traditional state party organizations, with their focus on transactional politics, have been displaced by a “volunteer” model that attracts ideologically motivated amateurs; the result, they find, has been to amplify the voices of ideological activists at the expense of rank-and-file voters. “The subset of voters who show up for primaries is more polarized than the general electorate,” write Bruce Cain and Cody Gray. As a result, they observe, “One unintended result of this shift [toward nomination by plebiscite] is that it has enhanced the activists’ more ideological voice in the primaries.” Political theorists sometimes refer to this empowerment of activists and ideologues as the problem of unrepresentative participation. Whatever one calls it, it has a perverse consequence: Parties that disempower their professional class become less representative of their rank and file.

No wonder political activists dislike professional vetting: It weakens their influence. Henry Jones Ford’s century-old insight is correct: Primaries do not disenfranchise elites; rather, they shift power from one set of elites (insiders who serve the party organizations) to another set (interest group activists and media personalities). By contrast, the public is quite comfortable with quality control. In a survey, one of us (La Raja) asked a random sample of voters to allocate 100 points to four different groups based on who they think should have influence in primary nominations. On average, respondents allocated 43% of the influence to “party voters,” and 21% of the influence to “independent voters.” But they allocated 19% of the desired influence to “party officials,” and another 16% to “nonpartisan experts.” In other words, survey respondents wanted to give professionals (party leaders and experts) more than a third of the influence in picking the nominee. Moreover, only 14% of respondents wanted to allocate all influence to voters.

In short, Americans view a mixed system as a good idea. To judge by the survey result, they want electorates to do about two-thirds of the deciding and parties and professionals to do about one-third, a ratio which strikes us as quite reasonable.

A further indication that most voters, as opposed to activists, do not mind quality control is that in House races—unlike presidential races—the parties continue to shape and organize candidate fields, to the limited but still significant extent that they can. In 2018, the Democratic Congressional Campaign Committee worked aggressively to weed out weak and extreme candidates in swing districts. For example, the committee persuaded a weak candidate in California’s 21st district to withdraw, and it campaigned in Texas’s seventh district against a candidate the party deemed too liberal to win; in both districts, the Democrats succeeded in flipping Republican seats. Although some activists were unhappy with those interventions, the party’s regulars were delighted by the wins and more voters got candidates they could support. The result was to make the process more representative, not less so. Moreover, because the Democrats were able to win a majority in the U.S. House, quality control advanced the interests of the broader party coalition.



Our case so far has dwelt on the shortcomings of the plebiscitary nominating process. So, we ought to re-emphasize: We are not saying that primary elections bring nothing to the table. To the contrary, they surface all kinds of important information about candidates and voters. What we do believe is that two filters are better than one. Electoral and professional perspectives check each other’s excesses and balance each other’s viewpoints; and, more than that, they complement and improve each other. Each provides the other with vital information which otherwise might be missed. Perhaps most important, professional input aids in winnowing the field to those who will likely govern competently.

“[T]wo filters are better than one. Electoral and professional perspectives check each other’s excesses and balance each other’s viewpoints[.]”

Insiders look for whether candidates are able to work with others, and whether they have sound judgment, adaptability, a nuanced way of dealing with problems, and influential relationships inside and outside government. Insiders also observe candidates’ character, and they can detect personal flaws that might affect sound decision-making. Insiders know from experience the attributes and talents necessary for effective governing. Voters are not privy to that kind of detailed, hands-on knowledge.

Moreover, even when they are equipped with good information, voters tend to undervalue governing skill and political relationships. Most voters, influenced by relentless media attention on the presidency and on confrontational politics, have a distorted sense of the powers of the president. Many subscribe to what Brendan Nyhan calls the Green Lantern theory of the presidency: The office’s super powers and the president’s sheer willpower can overcome obstacles to governance. Experienced professionals know better.

Professionals also have an eye out for leaders who speak to multiple constituencies. The choices made by professionals are less susceptible to hijacking by organized minorities and fringe elements, by celebrities and media manipulators, and by the inherent randomness of sequential elections in which anything could happen. Moreover, despite their warts, smoke-filled rooms and invisible primaries did a good job of finding qualified people who could unify their own party and exert broad appeal in a general election.

To be sure, the smoke-filled room did not always get it right. Warren Harding and Richard Nixon were poor choices (though both looked better in prospect than they turned out to be). That is why it is so important to have two filters. In the pre-1972 mixed system, primaries were not dispositive, but they showed party leaders which candidates were viable. A good example was John Kennedy. Beyond his popularity with Catholics, he needed to demonstrate his ability to appeal to Protestants, which he did when he won the West Virginia primary. Contrary to their popular image, smoke-filled rooms in the pre-1972 era did not simply impose their will on the party base; the system combined popular and professional elements, and candidates could and did successfully run as insurgents—Barry Goldwater being an obvious example. But even the maverick Arizona senator cut a formidable figure within his party. “Whichever path a candidate took,” notes Richard Pildes, “this system combined populist and party-centered features.”

“Vetting not only evaluates politicians; it also helps equip them to govern.”

By forcing candidates through a vetting process, professionals not only evaluate them and exact promises of accountability but also strengthen and deepen their ability to govern. In our system, political influence comes primarily from the “soft power” of relationships and political debts. By encouraging aspirants to build support among party leaders and elected officials throughout the country, vetting encourages candidates to cultivate connections and political capital, often long before they officially declare their candidacies. Vetting not only evaluates politicians; it also helps equip them to govern.


The professional filter also helps exclude candidates who are downright dangerous. Stephen Levitsky and Daniel Ziblatt, in their recent book, “How Democracies Die,” raise this point forcefully. In a democracy, they argue, party organizations’ most critical function is to act as gatekeepers against demagogues and charlatans. When parties lose or abandon that function, no other institution can dependably provide it. Levitsky and Ziblatt affirm what America’s Founders well understood: Democracies are typically destroyed not by violent overthrow but slowly, from within. The Founders were keenly concerned about the mischief of politicians with “talents for low intrigue, and the little arts of popularity,” as Alexander Hamilton famously wrote in Federalist No. 68, and they counted on vetting by the Electoral College to provide a failsafe. When populism reigns unchecked, or when establishment figures assume that potential authoritarians can be controlled in office, the results are frequently catastrophic. Hitler and Mussolini are paradigmatic cases, and Levitsky and Ziblatt offer numerous examples in recent times: Victor Orban (Hungary), Hugo Chavez (Venezuela), Alberto Fujimori (Peru), and Jaroslaw Kaczynski (Poland), among others. Those leaders and other authoritarians undermine democracy by changing the democratic rules and sidelining constitutional rivals who might challenge them.

It is Levitsky and Ziblatt’s view—and ours—that the Republican party failed in its gatekeeping role by permitting Trump to ascend to the nomination. He possesses obvious authoritarian tendencies and undercuts democratic norms at every turn. The institutional party should have had the capacity to keep him and candidates like him off the party ballot. But by the time Trump came along, the party establishment was a paper tiger; in 2012, the GOP seemed unable to refuse a platform to marginal candidates like Michelle Bachmann and Herman Cain. The door was ajar, and Trump pushed his way through it. The same door remains open so long as professional judgment is sidelined.

“[B]y the time Trump came along, the party establishment was a paper tiger[.]”

It was not always thus; parties have blocked anti-democratic candidates in the past. In 1924, Henry Ford, the legendary carmaker, considered running for president. He spouted anti-Semitic and racist commentary and promised to rid the nation of its corrupt and ineffective politics, claiming to be someone “who could do things and do them quick.” In both respects, he was an antecedent of Trump. Although Ford gained popularity, he did not stand a chance in either the Democratic or Republican party, because political professionals did not give him the time of day. He did not even bother going through with a candidacy, knowing he would be stymied by party leaders.

Or consider a more recent example: In 1976, Democrats in Florida were desperate to prevent former Alabama Gov. George Wallace from winning the state, as he had done in 1972, when six other candidates had split the anti-Wallace vote. A populist who had built his career on racism, Wallace was toxic to the party’s base outside of the South and likely unacceptable to the general electorate, and so his candidacy had the makings of a disaster. To stop him, party operatives leaned on other Democratic candidates to clear the field in Florida’s primary for Jimmy Carter, the only other southerner in contention. Sure enough, with the field narrowed, Carter defeated Wallace in Florida and ended Wallace’s candidacy (and went on to the White House—which the Democratic operatives had not expected). Still more recently, in 1996, DNC chair Don Fowler was able to exclude the conspiracy-mongering Lyndon LaRouche from the Democratic Party’s nominating process. The point is not that Wallace or LaRouche would have won the nomination had party leaders not intervened, or that the party intended to install Carter as the nominee (it manifestly did not). Rather, by intervening, the party protected the integrity of its brand and upheld its prerogative to set limits on who can run under its banner—a prerogative which is foundational to the very existence of a party as a meaningful political entity.

Open doors are an invitation to extremists and opportunists, but just as worrisome is renegade behavior by ordinary politicians—not only in their campaigns, but also in office afterwards. Officeholders respond to incentives. If tweeting belligerently, torpedoing compromises, and trashing democratic norms help them, then they will engage in those behaviors. If being team players, de-escalating conflict, and building effective coalitions help them, then they will engage in those behaviors.

In politics, both independence and accountability—both conflict and compromise—are important; the trick is to get the balance right, which requires using a mix of incentives. Today, however, voters in primaries lean toward combativeness and amateurism over compromise and professionalism. The presidential primary system selects for performance skills and performative behavior, more than for governing skills and constructive behavior. Predictably, the effects of voters’ anti-governing bias do not end on Election Day: Encouraging and training candidates to think of themselves as performance artists complicates governing day in and day out.

In their efforts to screen out renegades and incompetents, would professionals also screen out new ideas and overlooked constituencies? It is always a risk. Many observers who are dismayed by the Trump phenomenon acknowledge that it gave voice to working-class whites and victims of globalization whom mainstream Republicans (and mainstream Democrats) had neglected. But we decline to be forced into what we believe is a false choice between openness and competence. Down through American history, including the period when bosses had far more power than today, both parties have endured by taking on board popular grievances and insurgent movements, and they have proved nothing if not adaptable in their never-ending quest to form majority coalitions. Though professionals are hardly perfect, when they do their jobs, they do not exclude new claims and claimants (a self-defeating venture for a party that seeks to win); but they do try to channel them, bring them into the party, and reconcile them with prior claimants and with the imperatives of governing. A stronger GOP establishment in 2016 might have consolidated the candidate field to block Trump, but it also might have drawn on his energy and ideas in crafting its message and platform. And, indeed, the Democrats have done something very much like that in the current cycle: It was the defeated 2016 primary candidate, Bernie Sanders, rather than the nominee, whose ideas and style framed the party’s dominant agenda in the 2020 cycle.


The conventional assumption that primaries are less elite (whatever that means) than party selection overlooks the way today’s primaries actually work. Thanks to court decisions such as v. Federal Election Commission, there is today no limit on the size of contributions to independent groups; the groups, in turn, are free to support and oppose candidates provided that they not coordinate their activities with the candidates and parties. In other words, today’s campaign-finance rules funnel vast sums of unaccountable money to the political system’s least accountable actors.

“[T]oday’s campaign-finance rules funnel vast sums of unaccountable money to the political system’s least accountable actors.”

Those perverse rules have perverse consequences. Formerly compelled to seek funds from many establishment donors, candidates can now be bankrolled by quirky billionaires with pet agendas. In the 2012 Republican nomination cycle, for example, former House speaker Newt Gingrich’s campaign was kept alive by millions of dollars spent by Sheldon Adelson, a billionaire casino magnate, even though Gingrich’s support was weak among both the public and the party establishment. The infusion of funds from billionaire super PACs allows single-issue or fringe candidates to keep their ideas on the agenda, making it more difficult for the party to unify and prepare for the general election. Also, of course, billionaires can bankroll themselves, buying their way around accountability to any party or constituency. In October 2019, the hedge-fund billionaire Tom Steyer effectively bought his way into a Democratic candidate debate, spending about $30 million to generate $2 million in small donations—not an encouraging development if you think governing requires skill and experience. Michael Bloomberg could spend hundreds of millions of dollars on his Democratic campaign and consider it pocket change.

To be sure, there is a new alternative to billionaire financiers. The internet and other advances in technology make attracting small donors as easy as sending an email and soliciting a couple of clicks. Like many other analysts, we value the participatory enthusiasm of small donors; unlike many others, however, we acknowledge a troubling downside: In its current form, the small-donor revolution weakens the role of party gatekeepers and empowers fringe candidates. Academic research suggests that, far from being representative of the American electorate on a range of characteristics, small donors are as extreme and polarized as large donors, perhaps more so. Additionally, extremist candidates tend to do better at raising small-donor money, because they get the media’s attention by staking out bold (if unrealistic) positions and making attention-grabbing statements that often violate political norms. It is no coincidence that in the 2016 elections the disruptive populists in both parties, Bernie Sanders and Donald Trump, received more money from small donors than did any of the establishment-minded candidates. In fact, Trump shattered records, breaking Obama’s 2012 record with small donors, and accumulating slightly more than both Clinton and Sanders combined.

That said, even if small donors were a perfectly representative group, they would still provide a pathway around gatekeepers, and that is a mixed blessing. True, candidates who rely on small donors are less beholden to big donors and special interests, which may make them more independent-minded; also true, they are less beholden to their political peers, party leaders, and important constituencies, which may make them more reckless and demagogic.

Our point is not that small donations are necessarily bad (or good). It is that small donations are safer and more constructive in a system which provides professional vetting than in a free-for-all. Small donors (like voters) and political professionals have different job descriptions. They are not substitutes; they are complements. The nominating process is both more broadly representative and more likely to produce successful governance when amateurs and professionals collaborate.

Then there are the media, whose power in influencing candidate choice has grown enormously since the McGovern-Fraser reforms. Writing as long ago as 1978, Jeanne Kirkpatrick tartly observed:

“Advocates of the direct primary intended to wrest control from the bosses and return in to the people; presumably they did not intend to vest power in Walter Cronkite and other media moguls or to speed the development of a personalist politics with standards and practices more relevant to entertainment than to public affairs.”

Things have only gotten worse in the transition from Walter Cronkite to Sean Hannity and today’s bevy of extremist internet sites.

Why should anyone care if the media have influence—and if the most extreme media voices have the most influence? Media elites face completely different incentives than political professionals when they evaluate candidates. The media prefer the novel, the colorful, and the combative, qualities which drive compelling narratives. The problem, of course, is that those are not the same qualities which make for effective governing. Also, horserace coverage elevates the importance of early primary states, because it builds narratives around random swings in polling, unusual events, and candidate gaffes—all of which advantage candidates and consultants who are deft in the arts of spin, theatrics, and symbolic politics. Unlike party professionals, media figures need not think ahead about what happens after the ballots are counted, because they are not accountable for governing.

“Whether we consider access to money, to media, or to the ballot, cutting political professionals out of the nominating process makes the system less representative, less accountable, less competent, and thus less democratic.”

Social media, of course, provide candidates with pathways around traditional media elites. In 2020, however, it goes without saying that social media provide no serious vetting for governing skill; if anything, they are even more addicted to outrage, conflict, and emotional narratives than are traditional media—but without the guardrails against fake news and trolling which traditional media at least try to provide.

No amount of media democratization can substitute for professional judgment. In fact, without professional judgment, media democratization is more of a curse than a blessing. The point can be generalized: Whether we consider access to money, to media, or to the ballot, cutting political professionals out of the nominating process makes the system less representative, less accountable, less competent, and thus less democratic.



Most of today’s political reformers put their money on proposals to tinker with voting protocols (ranked-choice voting, proportional representation, multi-member districts), or to increase participation (mail voting, weekend voting, automatic registration), or to improve fairness (redistricting commissions, small-donor matches, donor transparency). Of those proposals, the most relevant to the problems we identify is probably ranked choice voting. In principle, it provides deeper information about voter preferences and may select candidates who are more satisfactory to a majority of voters, because it allows voters to register their second- and third-choice preferences. If no candidate receives a majority of votes on the first ballot, the last-place candidate is dropped. Her voters get to register their second choice on the subsequent (instantaneous) ballot. The process is repeated until a majority winner emerges. How ranked-choice voting would ultimately play out is hard to foresee. It might encourage coalition-building and help prevent the anointment of factional candidates. On the other hand, it makes voting more complex and cognitively demanding, and it might attract insurgent bids, no-hope candidates, and splinter parties seeking influence by running for second or third place.

Ranked-choice voting deserves to be tried, as do some other reform measures that focus on improving participation and equity. Our remit here is not to examine those proposals individually but to make a larger point about them as a class: Most process reforms are more likely to succeed as complements to professional input than as substitutes for it. No mechanical changes in the electoral system can substitute for rational parties and professionals in evaluating and organizing the candidate field. With so many candidates, so much strategic uncertainty, and so much confusing information, primary voters cannot reliably evaluate or organize the field by themselves, even if they were inclined to try (which they are not) and even if the system were optimally designed (which it is not). In primaries, where they are unable to use party labels to guide their choices, voters are especially prone to rely on momentary feelings, vague impressions, misleading rhetoric, fleeting events, and false information. Often, they cast their vote in protest, deliberately favoring self-expression and disruption over concern about governing.

“Most process reforms are more likely to succeed as complements to professional input than as substitutes for it.”

Mind you, nothing is wrong with voters using their gut to pick candidates or their ballot to protest. To the contrary, providing gut-checks and registering protest is the voters’ job. Professionals’ job is to assess candidates’ skill at uniting the party, building coalitions, and governing responsibly. It is up to professionals to nudge candidates to run for (say) a badly needed Senate seat rather than take a long shot at the presidency; it is up to professionals to consider how a candidate might fare among constituencies who are underrepresented in primaries but may be decisive in the general election (and in governing after the election); it is up to professionals to see that no one party faction can overwhelm and exclude others; it is up to professionals to deter renegade and antidemocratic behavior. Mixed systems ensure that the full spectrum of democratic values gets attention. They ensure a better balance between the democratic input of participation and democratic outputs of representation and governing. That is true regardless of the voting process and financing system used.

In our view, today’s most pressing—and most overlooked—task for political reform is not to perfect the populist filter but to restore and renew the professional filter. The goal should be to maintain the public’s voice in the selection of the party nominee, but to supplement it with the opinion of people who have experience in politics and governing. Although it is too late to make changes for 2020, both major parties should begin designing reforms that encourage more peer review.

How might they do so? Consider some possibilities—not necessarily as conclusive answers, but as examples.

Enhance the role of superdelegates

We recommend doing the opposite of what the DNC chose to do with its Unity Reform Commission. Instead of diminishing the role of superdelegates by preventing them from voting on the first ballot or reducing their numbers, the party should augment their influence.

The purpose of superdelegates has never been to overturn the choice of voters in primaries. True, in principle they might act as a last barrier to a manifestly unacceptable candidate, like George Wallace or Henry Ford—but even that is unlikely, if a candidate has won a decisive victory in the primaries. Rather, their real importance (and the reason Bernie Sanders and his followers crusaded against them) is their indirect influence on the upstream end of the process. Their convention votes incentivize candidates to reach out to them in the early stages of campaigns. A candidate who seeks superdelegates’ support will need to listen to them and promise to work with them. Also, superdelegates’ commitments early in the process help establish party support and momentum for favored candidates. Superdelegates do not decide the nomination, but they do influence the nominees, the media, and the voters—and that is exactly as it should be.

“We recommend doing the opposite of what the DNC chose to do with its Unity Reform Commission. Instead of diminishing the role of superdelegates …, the party should augment their influence.”

To some, the very thought of regarding professionals as a constituency seems undemocratic. But the elected officials who make up the majority of superdelegates represent voters in the party, a point which is usually forgotten amid populist moralizing. By virtue of standing for office and winning votes, elected officials are accountable to relatively broad public constituencies who are loyal to the party. The unpledged superdelegates are also party insiders who participate at all levels of government and have deep connections to diverse constituencies in the states. Politicians and insiders operating as unpledged delegates have incentives to strike a balance between activists’ policy demands and the need to assemble a broad (and winning) coalition. Pledged delegates, by contrast, represent the views of people who voted for a particular candidate; they are specifically bound to ignore or override the views of other voters in their states. In short, superdelegates think about coalitional necessities; pledged delegates, about winning a factional battle. Both perspectives are important, and neither is “undemocratic.”

A pre-primary vote of confidence

Another route, suggested by our Brookings colleague Elaine Kamarck, would be for authorized party leaders—members of Congress, governors, and party officials—to meet with presidential candidates before the first caucus and primary, interviewing candidates in private and then deciding whether to issue a vote of confidence or no confidence. Such a mechanism lets voters know which candidates are above the bar in terms of competence and adherence to party norms. The members of this party committee could vote for more than one candidate. Those not receiving a minimum number of votes (perhaps 15%) would be denied privileges afforded to other candidates, particularly appearances in national debates. Even if candidates below the threshold still appeared on the ballot, voters would have a clear signal from party leaders about the capabilities, character, and partisan loyalty of every contender.

Another, higher-tech method could take advantage of the sort of scoring metric seen at movie-review sites such as Metacritic and Rotten Tomatoes. Members of the national and state party committees would rate the qualities of announced candidates based on selected measures (perhaps, for example, political experience, ability to work with others, ethics, and so on). Scores would be tabulated and posted on party websites before the first caucus and primary. This approach avoids having to endorse just one candidate, while providing signals about candidates’ competence and party support, which would also be reported in the news media. The majority of scorers would be elected officials who represent voting constituencies (legislators, mayors, councilors). But they might also include a robust network of state and local party officials, which would enhance the value of working for the party committees (many of which are moribund). The candidate quality scores would encourage the news media to focus on metrics that matter for governing, and it would encourage the candidates to demonstrate their competence to co-partisans and the public.

A more formal kind of early vetting, at the state level, might require candidates to obtain petition signatures from a certain percentage of state and county party chairs and elected officials. Today, states routinely require candidates to obtain signatures from voters in order to reach the ballot; requiring signatures by party leaders and office holders would add the important second filter. There is no reason every state would need to adopt the same threshold or include the same assortment of leaders; nor is there reason to expect party officials to block candidates with grassroots support and prospects of winning (the backlash would be furious). Requiring a show of party support would, however, force candidates to make their case to party grandees—another way to signal acceptability and discourage rogue behavior.

Consider candidates’ track record and party service in allocating debate slots

For a variety of legal and political reasons, the parties’ authority over their own debates is constrained. Yet debates are very important for introducing voters to the party’s candidates. They are an essential aspect of the winnowing process. Selecting invitees is particularly challenging when the candidate field is large, as became evident in the Republican nominating cycle four years ago, when the candidates were so numerous that those who fell below a national poll threshold of 3.5% had to attend an “undercard” debate instead of the main attraction. One consequence was to favor a reality-television celebrity (Trump) over veterans like Sen. Lindsey Graham, an expert on foreign affairs who had served South Carolina in the Congress since 1993. That seemed shortsighted and unreasonable at the time, and it seems all the more so in hindsight.

Struggling with the same problem in the current nominating cycle, and with even more candidates jockeying to participate, the Democrats opted for a two-part participation test: Candidates needed to show they could attract a certain number of small donors, or they needed to garner poll ratings above a certain threshold, or, if the number of candidates exceeded 20, participants were required to meet both criteria. Those criteria glaringly excluded any reflection of candidates’ track record either within the party or in government. It would be perfectly reasonable, for example, for the party to use service in elective office as a consideration, or to consider the number and prominence of offices held, or to include input from superdelegates or state and local party chairs, or to give weight to major endorsements, or to prefer candidates who have campaigned and raised money for others in the party. By contrast, considering only polling numbers and small donations can put a thumb on the scale for media celebrities and demagogues, surely an instance of populism in its most perverse form.

In principle, there are limitless ways to empower party professionals. Abolishing or dramatically increasing contribution limits to the political parties, for instance, would give more control over the money taps to party insiders (and less to unaccountable outside groups). Or states could bypass the primary process altogether by using party regulars to name unbound convention delegates, which is what Colorado did in 2016. There is no need for every state to make the same change, so long as insiders get a voice. Nor, by the same token, is there any one “right” reform. When the smoke-filled room and the invisible primary worked best, they exerted influence through many channels, formal and informal. The professional filter worked because it developed organically and included many voices and views. The right solutions, now as then, are the ones which the parties can implement, and which get the job done.


The mechanics of peer review are almost trivially easy. What is difficult is the politics. The parties fear the kind of excoriation that Bernie Sanders’s supporters unleashed against superdelegates. They worry about being blamed for unpopular decisions. Far easier to pass the buck to the voters, allowing the parties to become vehicles for whoever survives (or hijacks) the primaries. The parties’ and professionals’ confidence in their own legitimacy and efficacy has collapsed—a self-fulfilling prophecy, and, as it turns out, a dangerous one. Under the circumstances, it is sad but not surprising that the Democratic Party has moved to weaken superdelegates.

The most important reform, then, is to change the mindset that regards popular elections as the only acceptable way to choose nominees. No democratic nation has opened party leadership selection to the same degree as the United States, and recent experiments with primaries by Britain’s Labour Party and France’s Republican Party and Socialist Party have proved disastrous for all three parties, empowering extremists and smashing party coalitions. (A European diplomat told us of France, “You can make a case that the primaries destroyed both of the mainstream parties.”) Paradoxically, democracy fundamentalism—the insistence that the remedy for whatever ails democracy must be more democracy—is dangerously undemocratic, as America’s Founders well understood. And political consumerism—the idea that more choice is always better—is a recipe for chaos.

A healthier approach prizes institutional actors. Only they can provide expertise on the qualities and trade-offs necessary for governance; only they can speak for the long-term interests of the party. Their institutional positions and instincts for self-preservation lead them to understand the party not simply as a vehicle for one candidate’s ambitions, but as a national political association with a past, present, and future, and with responsibilities for governing, not just winning.

“Today’s task is not to banish populism (as if that were possible) but to balance it, restoring the Madisonian pillars of pluralism, checks on power, and deliberative institutions.”

Our arguments for providing for peer-review and limiting choice will be met with scorn by populists who think peer review denudes democracy. In truth, however, peer review seeks to assure that democracy continues to thrive. Today’s task is not to banish populism (as if that were possible) but to balance it, restoring the Madisonian pillars of pluralism, checks on power, and deliberative institutions.

Can it be done? The public, as we have seen, is receptive to quality control. Even many democracy fundamentalists and political consumerists have come to see a problem when the nominating process is open to dozens of contenders, some of whom are more interested in building their brands than in winning, much less governing. Because peer review was the norm in American politics for all but the last decade or so of our political history, no one can say it is a radical, alien, or untried scheme. Also, of course, the Republican nomination’s capture by an ignorant and illiberal candidate (who was not even, in any meaningful sense, a Republican) has been a wake-up call. Without peer review, America’s nominating system will remain a chaos process giving birth to chaos candidacies.

The authors are grateful for comments and inspiration from Bob Bauer, Bruce Cain, David Hopkins, John Hudak, Elaine Kamarck, and Richard Pildes.

This work is licensed under the Creative Commons Attribution-NonCommerical-NoDerivatives 4.0 International License. To view a copy of the license, visit


  • Footnotes
    1. Primaries for other offices, including seats in Congress and governorships, are prone to many of the same problems. Many of our arguments apply to non-presidential races, but we limit our scope here to the presidential level.
    2. For a brief history of the disjointed development of the nomination process and an illuminating comparative perspective on how parties in other democracies choose candidates, see Stephen Gardbaum and Richard H. Pildes, “Populism and Institutional Design: Methods of Selecting Candidates for Chief Executive Symposium: A Debatable Role in the Process,” New York University Law Review 93 (2018), 647-708.
    3. Gardbaum and Pildes, ibid. France’s Socialist Party started using primaries to pick its presidential nominee in 2007, while the Republican Party began using them for the 2017 election.
    4. Byron E. Shafer, Quiet Revolution: The Struggle for the Democratic Party and the Shaping of Post-Reform Politics (Russell Sage Foundation, 1983).
    5. Hedrick Smith, “Democrats Back Carter on Nomination Rule: Kennedy Withdraws from Presidential Race,” The New York Times, August 12, 1980.
    6. See Elaine C. Kamarck, Primary Politics, third edition (Brookings, 2019), pp.62-67.
    7. Elaine C. Kamarck, “Returning Peer Review to the American Presidential Nominating Process” New York University Law Review 93 (2018), 709-727.
    8. This invisible primary has been an informal vetting process that precedes caucus and primary elections. Elected and non-elected party officials use candidate endorsements and their influence over the party’s major donors to whittle the field to candidates who are deemed competent, well connected, and loyal to the party. The rise of partisan media, social media, online donations, mega-donors, super PACs, and other deep-pocketed outside groups has eroded the influence of party leaders during the pre-primary period. For an extensive review of the invisible primary, see Kamarck (2019) and Cohen et al. (2008), cited elsewhere.
    9. Marty Cohen, David Karol, Hans Noel, and John Zaller, The Party Decides: Presidential Nominations Before and After Reform (University of Chicago Press, 2008). The authors argued that the “party was back” in control of the nomination process, exerting influence through endorsements and converging on favored candidates. In a critique of this argument, Christopher Achens and Larry Bartels pointed out that, in four of the past six nominating cycles, one party or the other had failed to consolidate its support. In 2004, for example, few governors endorsed anyone, and John Kerry, who won the Democratic nomination, did not have the most endorsements. In 2008, the GOP split among three candidates (John McCain, Rudy Giuliani, and Mitt Romney); Democratic endorsements split between Hillary Clinton and John Edwards, with Barack Obama getting just 9%. See Christopher H. Achen and Larry M. Bartels, Democracy for Realists: Why Elections Do Not Produce Responsive Government (Princeton University Press, 2016), p. 65. It is clear that in 2012, the GOP establishment successfully weighed in for Romney, and in 2016, the Democratic establishment did the same for Clinton, showing that party influence was not a dead letter. But more striking in 2016 was the Republican establishment’s complete failure, with Trump winning with practically no endorsements.
    10. Benjamin Wittes and Jonathan Rauch, “More Professionalism, Less Populism: How Voting Makes Us Stupid, and What to Do About It” (Brookings, May 27, 2017).
    11. Richard R. Lau, “Correct Voting in the 2008 U.S. Presidential Nominating Elections,” Political Behavior 35 (2013). For example, in the 2008 Republican primaries, Lau “calculate[s] that only 50 percent of Republican primary and caucus voters voted correctly.”
    12. Christopher H. Achen and Larry M. Bartels, Democracy for Realists: Why Elections Do Not Produce Responsive Government (Princeton University Press, 2016).
    13. Bruce E. Cain, “Populists Illusions and Pluralist Realities,” The American Interest 10:2 (October 3, 2014).
    14. Among post-Civil War presidents who won the presidency, nominees who won after multiple ballots included Hayes (seven ballots), Garfield (36), Cleveland (two), Harrison (eight), Wilson (46), Harding (10) and F. Roosevelt (four). See Drew Silver, “Contested Conventions in Presidential Races, 1868-1984,” Pew Research Center, February 3, 2016.
    15. The political scientist Nelson Polsby observes: “A premium is placed … in hoping that the field becomes crowded with rivals who cluster at some other part of the ideological spectrum, or who for some other reason manage to divide up into too-small pieces the natural constituencies that exist in the primary electorate.” See Nelson W. Polsby, Consequences of Party Reform (Oxford University Press, 1983), pp. 68-69.
    16. Samantha Smith, “Voters Have a Dim View of Primaries as a Good Way to Pick the Best Candidate,” Pew Research Center, April 5, 2016.
    17. Caroline J. Tolbert, David P. Redlawsk, and Daniel C. Bowen, “Reforming Presidential Nominations: Rotating State Primaries or a National Primary?” PS: Political Science & Politics 42:1 (2009), pp. 71–79.
    18. Barbara Norrander, The Imperfect Primary: Oddities, Biases and Strengths in U.S. Presidential Nomination Politics, second edition (Routledge, 2015).
    19. Henry Jones Ford, “The Direct Primary,” North American Review 190:644 (1909), pp. 2-4.
    20. Nelson W. Polsby, Consequences of Party Reform (Oxford University Press, 1983).
    21. Dennis C. Spies and André Kaiser, “Does the Mode of Candidate Selection Affect the Representativeness of Parties?” Party Politics 20:4 (2014). Another large-sample study arrives at similar findings, indicating that inclusiveness and participatory openness in the selection process may come at the cost of representativeness and responsiveness. See Reuven Hazan and Gideon Rahat, Democracy Within Parties: Candidate Selection Methods and Their Political Consequences (Oxford University Press, 2010).
    22. Raymond J. La Raja and Brian F. Schaffner, Campaign Finance and Political Polarization: When Purists Prevail (University of Michigan Press, 2015).
    23. Byron E. Shafer and Regina L. Wagner, The Long War Over Party Structure (Cambridge University Press, 2019).
    24. Bruce E. Cain and Cody Gray, “Parties by Design: Pluralist Party Reform in a Polarized Era,” New York University Law Review 93 (2018), pp. 636 and 632.
    25. This dynamic is especially strong in a crowded field, making it difficult for political elites to find consensus on candidates and communicate preferences through endorsements and other informal processes. When this happens, the power of the media to influence poll standings of candidates surges. See John Sides, Michael Tesler, and Lynn Vavreck, Identify Crisis: The 2016 Presidential Campaign and the Battle for the Meaning of America (Princeton: Princeton University Press, 2018).
    26. It may be noteworthy that the public believes that partisans should be accorded much more influence than independents in selecting party nominees, a preference which we believe to be well grounded. Zachary Albert and Raymond La Raja, “Who Should Decide the Party’s Nominee? Understanding Public Attitudes Toward Primary Elections.” Working paper, APSA Preprints (2019).
    27. Brendan Nyhan, “Donald Trump, the Green Lantern Candidate,” The New York Times, August 25, 2015.
    28. See Jamie L. Carson and Jason M. Roberts, Ambition, Competition, and Electoral Reform: The Politics of Congressional Elections Across Time (University of Michigan Press, 2013).
    29. Richard Pildes, “Two Myths about the Unruly American Primary System,” The Washington Post, May 25, 2016.
    30. See James W. Ceaser, Reforming the Reforms: A Critical Analysis of the Presidential Selection Process, 1982 (Ballinger Publishing 1982); and Steven Levitsky and Daniel Ziblatt, How Democracies Die (Penguin, 2018).
    31. Others have made this point, including Aziz Huq and Tom Ginsburg, “How to Lose a Constitutional Democracy,” UCLA Law Review 65 (2017).
    32. Levitsky and Ziblatt (2018).
    33. Levitsky and Ziblatt (2018), pp. 43-46.
    34. Jon Ward, Camelot’s End: Kennedy vs. Carter and the Fight that Broke the Democratic Party (Twelve, 2019), Kindle location 1311. Sen. Henry Jackson also remained in the Florida primary race and finished third.
    35. The eventual 2012 nominee, Mitt Romney, spent enormous time and resources trying to fend off these well-resourced factional candidates well into the spring before the convention. Fatefully, in response to this dynamic the RNC changed the rules for 2016 to allow delegates to accumulate more rapidly for one candidate, a dynamic that helped Trump.
    36. See, e.g., Bryan Anderson, “Did Tom Steyer Buy His Way into the Democratic Debate? How He Worked His Way to the Stage,” Sacramento Bee, October 15, 2019; Dan Merica, “Tom Steyer Raises Paltry $2 Million after Personally Spending Tens of Millions on His Presidential Bid,”, October 10, 2019.
    37. Adam Bonica and Jenny Shen, “How Wealthy Campaign Donors May Reduce Political Polarization and Weaken the Tea Party,” Washington Post, April 24, 2014.
    38. Bertram Johnson, “Individual Contributions: A Fundraising Advantage for the Ideologically Extreme?” American Politics Research 38:5 (2010), pp. 890–908.
    39. Small donors are those who gave a total of $200 or less to candidate. Trump raised $239 million from small donors (69% of his total individual contributions), compared to Sanders’s $137 million (44% of his total individual contributions), and Clinton’s $100 million (22% of her total individual contributions). Obama raised $219 million in small contributions in 2012. See Campaign Finance Institute, “President Trump, with RNC Help, Raised More Small Donor Money than President Obama; As Much As Clinton and Sanders Combined,” press release, February 21, 2017.
    40. The same is true of public financing. As Arizona has found, public funds are an entry ticket to extremists and outsiders who behave bizarrely in office. Public financing better accomplishes its goals in the context of a vetted nominating system.
    41. Jeanne J. Kirkpatrick, Dismantling the Parties: Reflections on Party Reform and Party Decomposition (American Enterprise Institute, 1978), p 22.
    42. We gratefully draw here upon the suggestions of our Brookings colleague Elaine Kamarck, who is an expert on primaries and a member of the Democratic National Committee. See, among her other important works on peer review in the nominating process: Primary Politics: Everything You Need to Know about How America Nominates Its Presidential Candidates (Brookings, 2018); “Re-Inserting Peer Review in the American Presidential Nomination Process” (Brookings policy brief, April 27, 2017); “Returning Peer Review to the American Presidential Nomination Process,” New York University Law Review 93:709 (2018).
    43. In 1981, when the party was trying to amend the populist reforms inspired by the McGovern-Fraser Commission, Rep. Geraldine Ferraro pointed out the obvious to the Hunt Commission: “No one knows those people [rank-and-file Democrats] better than a Member of Congress. No one is better able to represent them at a convention than a Member of Congress.” See “Hunting the Hunt Commission,” In These Times, May 16, 2016.
    44. See Bob Bauer, “A Debatable Role in the Process: Political Parties and the Candidate Debates in the Presidential Nominating Process,” New York University Law Review 93(4), 2018.
    45. Frances McCall Rosenbluth and Ian Shapiro, Responsible Parties: Saving Democracy from Itself (Yale University Press, 2018).
    46. See, e.g., “Here’s What to Expect from Colorado’s GOP Caucuses,” The Colorado Independent, January 5, 2016.
    47. See Richard H. Pildes, “Romanticizing Democracy, Political Fragmentation, and the Decline of American Government,” Yale Law Journal 124 (2015), 804-852.
    48. See Levitsky and Ziblatt (2018) on the vital role of political parties and their leadership as gatekeepers that preserve democracies.