How the Iowa caucus predicts presidential losers, not winners

While a slew of presidential candidates are hoping to win next week's caucus in Iowa, it's really more about not losing.

What do Jimmy Carter, Ronald Reagan, George H.W. Bush, Bill Clinton, Michael Dukakis, Mitt Romney and John McCain all have in common? All of these politicians went on to win their party’s presidential nomination after losing the Iowa caucus.

Since 1972, there have been nine Democratic and seven Republican contested caucuses. Only five of the Democratic caucus winners and three of the Republican caucus winners have gone on to win their party’s nomination. With success rates of 55 and 43 percent respectively, it’s clear the Iowa caucus isn’t great at predicting presidential primary winners. 

So why are candidates and media so obsessed with this corn-growing state of three million people?

One answer: Iowa might be less about deciding the winner, and more about confirming the losers. 

Since 1972, no Democratic or Republican candidate who finished worse than fourth place in Iowa has gone on to win their party’s nomination. And both parties have only had one case of fourth-place survival: Democrat Bill Clinton in 1992 and Republican John McCain in 2008. For the other ten caucuses during this time, the eventual party nomination was always awarded to a top-three candidate in the Iowa caucus.

“In the 1980 Republican presidential caucus campaign, Tennessee Sen. Howard Baker said one function of the Iowa caucuses was to ‘winnow the field’ of candidates,” Des Moines Register political columnist David Yepsen wrote in 2007. “By that he meant Iowa caucus-goers in both parties take presidential campaigns with large numbers of candidates and cut the field to a more manageable size for voters in other states to consider." 

So how does Iowa winnow the field? By confirming or defying expectations, say political observers. 

“Every candidate in Iowa has the same opponent, and that opponent’s name is ‘expected,’” Drake University political scientist Dennis Goldford tells Vox. “The caucuses are about who exceeds expectations and who fails to.” 

Before the 2004 Iowa caucus for example, Sen. John Kerry was third in the national polls behind Gov. Howard Dean with 25 percent support and Gen. Wesley Clark with 19 percent. After Sen. Kerry won an upset victory in Iowa, he rocketed to the top of the pack with a national preference of 47 percent, and later went on to win the Democratic Party’s nomination. 

And four years later before the 2008 election, all signs pointed to Hillary Clinton winning the Democratic nomination. In an Oct. 2007 Gallup poll, Clinton led President Barack Obama nationally 50 percent to 21 percent. But two months later, on Jan. 3, Obama won the Iowa caucus with 37.6 percent support, with Clinton in third place with 29.5 percent behind John Edwards.

“The results of Iowa were validating for us,” Larry Grisolano, a former campaign consultant for Obama tells Vox. “People became convinced that Obama was more than just a media phenomenon – and that he was a candidate who could attract votes.” 

But the Iowa caucus doesn’t help those who exceed expectations as much as it hurts those who fail expectations. After a loss in Iowa, voters, donors, and even candidates themselves doubt the campaign’s future prospects. 

“Until the voting begins, most candidates will probably still maintain some belief that it will all come together for them,” The New York Times’ Seth Masket writes in November. “In addition, with polling in such a state of uncertainty, some candidates will stick around until Iowa and New Hampshire make it painfully clear that a bid is over.” 

In short, Iowa can sometimes predict the winners, but it almost always predicts the losers.

In the 2008 caucus, the two lowest-scoring Democratic candidates dropped out of the presidential race within the week, and the Republican pool of candidates decreased from seven to three after a little more than a month. In 2012, the three lowest-scoring Republican candidates dropped out within a few weeks.