The presidential primaries are one of the most important elements of the American constitutional order. Given that general elections give voters just two starkly opposed choices, it's largely through the primaries that nuance enters the political process. Parties define themselves by whom they select to run for president, and the ideological alignments that result end up defining the contours of political conflict.
And yet, despite primaries' central role, nothing about them is laid out in the Constitution.
In fact, the framers didn't envision American politics taking the form of two-party competition, so they gave no thought to how parties would select their candidates.
This, in turn, is part of what makes the primaries so fascinating. While the Constitution itself is incredibly difficult to change, party nominating rules and state laws are much more flexible.
Consequently, the presidential nomination process is one of the elements of the American political system that's changed the most — and often in ways that aren't anticipated by the people driving the change.
Which leads to the last thing that makes primaries so fascinating: They are genuinely unpredictable. Conceivably almost anything could happen.
Early nomination contests didn't involve primaries
Presidents Andrew Jackson and Martin Van Buren. (Universal History Archive/UIG via Getty Images)
Intraparty disputes over who should be nominated for the presidency are as old as the republic itself. But the modern system of determining nominees through a series of state primary elections is essentially an innovation of the 1970s. Before that, parties deployed a wide range of methods.
The Democratic-Republicans, the dominant political party of the early 19th century, used to select candidates via a vote of the party's members in Congress. That method let it control the White House for 20 years, and lasted until the rivalry between John Quincy Adams and Andrew Jackson made the party splinter into the Democrats and the Whigs in the aftermath of the 1824 election.
Back in 1836, in the early days of Whig versus Democrat competition, the Whig Party even tried nominating several candidates simultaneously in their bid to block Martin Van Buren from succeeding Jackson in the White House.
In most Northern states, William Henry Harrison appeared on the general election ballot, while Hugh White got the nod in most Southern ones. And Massachusetts Whigs went with Daniel Webster (who carried the state), while Willie Magnum was nominated in South Carolina.
The idea was that running multiple candidates with distinct regional appeals could successfully deny Martin Van Buren a majority in the Electoral College, throwing the choice to the US House of Representatives. The selection of the Whig on each state's ballot was left up to the local party. Had the gambit worked, one could imagine the system of multiple nominees becoming entrenched.
But it did not work. Van Buren won the election, and in subsequent contests the Whigs emulated the Democrats, picking a single nominee at a broad national convention with representatives from all states.
Conventions are still held today, but they are essentially publicity stunts. At best, they're counting exercises in which the point is simply to crown the candidate who already enjoys the support of most of the delegates.
But historical conventions were real decision-making bodies, where a cast of locally selected elites would come together to genuinely choose someone. That opened the door to outcomes like the Whigs drafting celebrity war heroes Zachary Taylor and Winfield Scott in 1848 and 1852, without the candidates needing to mount vigorous primary campaigns.
The convention system also allowed for the emergence of "dark horse" candidates like 1844 Democratic nominee James K. Polk, who was not even considered a contender at the start of the process. He emerged on the convention floor as a broadly acceptable second choice after various factions deadlocked.
Conventions generally had delegates take a series of votes to winnow the field. At the 1860 Republican convention, for example, William Seward received by far the largest number of votes on the first ballot, with Abraham Lincoln finishing a rather distant second.
Seward's problem, however, was that virtually all of the delegates who weren't for him were strongly against him on electability grounds — as governor of New York he had not only opposed slavery, but also signed laws advancing the rights of free black residents of the state, radical moves that much of the party thought went too far for the swing states of the Midwest.
Lincoln's result gave him enormous momentum. On the second ballot, he trailed Seward by just three votes. And at that point, it just took a little more cajoling for Lincoln to get over the top.
When presidential primaries started, they weren't decisive
Woodrow Wilson. (Tony Essex/Hulton Archive/Getty Images)
The Progressive Era at the beginning of the 20th century saw a backlash against local party machines and their bosses dominating American politics. This backlash was especially pronounced in Western states, where reformers implemented ideas like legislating via ballot initiative at the polls.
Progressive reformers also invented the presidential primary. In 1910, Oregon became the first to use a popular election to pick its delegates for national conventions, with the delegates pledged to support specific candidates.
But these primaries lacked the efficacy and decisiveness of those we have today, in part because most states didn't have them and in part because the ultimate nomination decision was still made via a multi-ballot process at a national convention.
In 1912, ex-President Theodore Roosevelt decided to challenge his successor William Howard Taft for the GOP nomination. He crushed Taft in the primaries, carrying nine of the 12 states that held primaries, while Robert La Follette won two and Taft just one.
But that still left 36 other states, which mostly sent pro-Taft delegates to the convention, securing him the nomination. And that led Roosevelt to bolt the party and launch an independent bid for the general election.
That year's Democratic convention, meanwhile, required 24 rounds of balloting for Woodrow Wilson to prevail over the now-forgotten House Speaker Champ Clark. That meant that delegates' initial pledges to specific candidates were long irrelevant by the time the final decision to nominate Wilson was made.
Twelve years later, primary voters again found their preferences overridden when in 1924 William McAdoo swept the Democratic primaries (largely held in the South and West) with the strong support of the Ku Klux Klan. But for precisely that reason, he was totally unacceptable to the party establishment back East, whose machines relied on the loyalty of Catholic voters.
But while McAdoo didn't have enough support to win, he did have enough to block the party bosses' favorite, New York Gov. Al Smith, a Catholic.
After a brutal 99-ballot war of attrition at the Democratic convention, both McAdoo and Smith simultaneously withdrew. Then after "only" four more rounds of voting, a candidate nobody particularly liked, John W. Davis, got the nod — then got destroyed in the general election.
Primaries were beauty contests
Eisenhower (ullstein bild/Getty Images)
When primaries did play a substantive role, it was instead through their function as beauty contests. Winning the 1952 New Hampshire primary let Dwight Eisenhower prove that rank-and-file Republicans, and not just party bosses, were more interested in picking a winner than in picking an orthodox conservative — thus giving the establishment permission to do what it wanted and go with Ike.
By the same token, winning the West Virginia primary in 1960 was a way for John Kennedy to demonstrate to party leaders that a Catholic could win votes in the South.
But both of these examples were making a point to persuade party leaders, not a way to override their preferences.
The fundamental inefficacy of the primaries was driven home by the bitter 1968 Democratic nomination contest that ultimately went to Vice President Hubert Humphrey, who didn't even enter any primary elections.
But the tumultuous, riot-scarred convention where it happened, followed by electoral defeat at the hands of Richard Nixon, spurred massive change.
The '70s were a crazy time for everyone
Jimmy Carter (Hulton Archive/Getty Images)
After the 1968 fiasco, the Democratic National Committee created a commission charged with proposing reforms to the nominating process. (It was chaired initially by Sen. George McGovern and then by Rep. Donald Fraser.)
Its report brought state delegate allocations into line with the distribution of population and required state parties to adopt open procedures for selecting delegates rather than allowing state party leaders to pick them in secret.
In practice, states mostly implemented this by adopting presidential primaries — which generally induced Republicans to make the same change.
The new system kicked off a chaotic era in which mavericks and factional leaders could win over the objections of party leaders.
In 1972, McGovern took advantage of his own reforms to win the Democratic nomination, even with an ideology so unacceptable to major party factions that the AFL-CIO didn't support him over Richard Nixon.
Then in 1976, Jimmy Carter won the Democratic nomination despite a total lack of ties to the party establishment in Washington, and proceeded to win the White House and then not pursue the party's agenda.
Also in 1976, incumbent President Gerald Ford faced an extremely strong primary challenge from conservative leader Ronald Reagan and was forced to drop the incumbent vice president from the ticket in order to appease conservatives.
Four years later, incumbent President Carter was challenged from the left by Ted Kennedy, his renomination secured only by the rally-round-the-flag effect induced by the Iranian hostage crisis.
At around this time, it became fashionable to observe that American political parties were in decline. University of California Irvine political scientist Martin Wattenberg achieved the apogee of this literature with his 1985 classic The Decline of Political Parties in America (since updated in five subsequent editions), citing the waning influence of party professionals, the rise of single-issue pressure groups, and an attendant fall in voter turnout. After all, a party whose leaders can't even pick its own presidential nominee in a reliable way isn't much of a party at all.
Elites still matter enormously in primaries
George H.W. Bush (ABC Photo Archives/ABC via Getty Images)
Just when journalists and political scientists were ready to proclaim the death of parties in favor of candidate-centered politics, the pendulum started to swing back.
Over the past 35 years, incumbent presidents have had zero problems obtaining renomination — even presidents like George H.W. Bush and Bill Clinton who alienated substantial segments of the party base with ideological heterodoxy during their first term. Reagan and Clinton both passed the baton to their vice presidents without much trouble.
Insurgent candidates who caught fire with campaigns explicitly promising to shake up the party establishment — Gary Hart in 1984, Pat Robertson in 1988, Jerry Brown in 1992, Pat Buchanan in 1996, John McCain and Bill Bradley in 2000, Howard Dean in 2004, Mike Huckabee in 2008, and Rick Santorum in 2012 — repeatedly gained headlines and even won state primaries.
But while 1970s insurgents were able to use early wins to build momentum, post-Reagan insurgents were ground down by the sheer duration and expansiveness of primary campaigns.
Tactics that worked in relatively low-population, cheap states like Iowa and New Hampshire simply couldn't scale without access to the broad networks of donors, campaign staff, and policy experts that establishment-backed candidates enjoyed.
In their 2008 book The Party Decides: Presidential Nominations Before and After Reform, Marty Cohen, David Karol, Hans Noel, and John Zaller argued that the resurgence of establishment candidates was no coincidence.
They argue that party insiders had found a way to control nominations by replacing the old smoke-filled rooms of the convention with a new series of insider bargains largely struck before convention voting begins.
It's this "invisible primary" among party elites that truly matters.
To test their idea, the book's authors tallied up endorsements from a broad set of party figures across two and a half decades of primaries. They included everyone from famous elected officials to local politicians to activists to celebrities, and calculated each endorser's importance in the party.
If one candidate was the clear winner in pre-Iowa endorsements and also won the nomination, then it could be said that the party had decided.
And that's just what they found. In eight of 10 competitive presidential primary contests between 1980 and 2004, endorsements showed that party insiders clearly backed one candidate before Iowa, and that candidate then went on to win the nomination.
Endorsements were better at predicting the outcome than polls, fundraising numbers, or media coverage.
The authors don't argue that endorsements alone specifically cause a candidate to win. Rather, endorsements are a signifier of how the invisible primary is going — and therefore of which candidate the party network is choosing to favor.
"In our theory, party insiders rally to the candidate of their choice, endowing him or her with endorsements, access to fund-raising networks, and pools of talent and volunteer labor," they write.
It's a theory that gained enormous prestige during the 2012 Republican primary cycle, which saw a series of novelty candidates rocket and then tumble in the polls, only for Republican voters to eventually settle on Mitt Romney, whom the establishment had favored the whole time.
Romney's relatively calm affect, his moderate record as governor of Massachusetts, and his Mormon faith all gave him trouble connecting with the conservative grassroots. But in the end, it didn't matter any more than it mattered for Taft or Humphrey or the forgotten Davis — the party thought he offered the best combination of commitment to conservative principles and electability, so he got the nod.
2016 could prove everything wrong
Donald Trump (MANDEL NGAN/AFP/Getty Images)
The problem with trying to understand the rules governing presidential elections is that there simply aren't very many.
A 35-year streak without a non-party-leaders-approved McGovern-type scenario or a strong challenge to an incumbent president is pretty striking. But it's certainly not out of the realm of possibility that it was a coincidence.
Back in 2004, Democrats ended up rejecting the Howard Dean insurgency, but the party united around Kerry as an alternative only after the Iowa caucuses — not before.
And in 2008, Obama didn't run as a true anti-establishment outsider and had plenty of party insider support. But he was certainly an underdog vis-à-vis Hillary Clinton in terms of endorsements and party connections.
If you squint at these elections right, you can definitely make them fit the model of elite-driven candidate selection. But if Ted Cruz manages to build on his Iowa caucus win to secure the Republican nomination — or if Donald Trump manages to hold on to his national polling lead — we'll likely look back on those years as early tremors that warned of a larger crack-up.
Contested primaries are a long, hard slog
Through the 1960s, nominations were typically wrapped up in the summertime — at the convention itself. Modern elections, by contrast, frequently give us primary races that wrap up within weeks of the New Hampshire primary in February, even though active campaigning may take place for a full year leading up to New Hampshire.
In 2004, for example, John Kerry rocketed out of nowhere to a big national lead after his unexpected win in Iowa in late January. And by March 2, all his opponents had dropped out.
But the long slogs of the 1970s and the brutal Obama-Clinton battle of 2008 serve as a reminder that the system as it exists on paper calls for a very long series of primaries.
New Hampshire, Nevada, and South Carolina all vote in February, followed by a large group of primaries on March 1. But though these early states are important for establishing candidates' viability and shaping media narratives, they still leave out huge piles of delegates to be won on March 15 (Florida, Illinois, and Ohio), April 19 (New York), and even June 7, in the very late California and New Jersey primaries.
It would be very unusual for a race to still be meaningfully contested that late, but a lot about the 2016 cycle thus far has been unusual. And the only real constant in the American candidate selection system is that it's always changing — often in unexpected ways.