The Iowa caucuses could be messy. On Monday night, for the first time, there will be four sets of results, and there is no guarantee that the same candidate will win them all.
Four different candidates could even declare victory, with the right combination of votes, turnout, geography and luck.
We’ve decided to show how something like that perfect storm could happen. This scenario is made up, but it’s entirely plausible and grounded in the data collected in the most recent New York Times/Siena College survey. Although it’s not likely, this hypothetical is still illustrative.
The first alignment
The first measure of the Iowa result is the “first alignment” or “first preference,” as it reflects the initial preference of caucusgoers when they arrive at their precinct. The preference for Iowa voters at this stage has never been publicly available before this year.
This is the result that ought to be closest to the pre-election polls. Let’s suppose that the first preference happens to come close to the results of recent polls, with Bernie Sanders in the lead, as in the chart below:
The final alignment
The second measure is called the “second alignment or “final alignment,” and it reflects the preference of caucusgoers after the supporters of nonviable candidates have an opportunity to realign with candidates who remain viable in their precinct. Usually, a candidate needs 15 percent in a precinct to be viable.
It’s easy to imagine how the results at second alignment could be quite different from the first.
The candidate who led on first alignment could fall behind under a few conditions. That candidate’s support could be geographically narrower, leading the candidate to fall beneath viability in more precincts than the nearest rival. The candidate might also fare worse among the former supporters of nonviable candidates, allowing another candidate to leap ahead.
To simulate second alignment for our perfect storm, we’ve estimated the first preference and second choice of every registered voter in Iowa using the most recent Times/Siena poll. This is not a prediction and, indeed, we made a couple of tweaks to bring about a perfect storm. But all of this is plausible, and very close to our actual data.
On second alignment, let’s imagine the statewide vote flips: Now Joe Biden leads.
Mr. Biden advances on second preference in our hypothetical for two reasons. First, his support is geographically broad. He clears viability in 90 percent of precincts, allowing him to keep an overwhelming share of his support from first alignment. Mr. Sanders has somewhat narrower geographic support, and clears viability in 80 percent of precincts. He falls short in many vote-heavy suburban areas.
Mr. Biden also benefits from the geographic distribution of the support of second-tier candidates. In the hypothetical, Amy Klobuchar falls short of viability nearly everywhere, and her support breaks to Mr. Biden.
Elizabeth Warren, on the other hand, remains viable in precincts containing around half of the vote, denying Mr. Sanders his most obvious source of upside. Where viable, she does well on second alignment — keeping her vote share intact over all.
Worse still for Mr. Sanders, she remains viable in the more liberal areas where her backers are likeliest to prefer him as their second choice, like on college campuses. She falls short of viability in older, less educated and more conservative areas where her voters don’t overwhelmingly back Mr. Sanders on second choice.
State delegate equivalency
Third, each precinct caucus selects delegates to county conventions, based on the results at final alignment. The results are translated into estimates of how many delegates each candidate will earn at the state convention, what’s known as “state delegate equivalents.”
The candidate with the most state delegate equivalents is traditionally the winner of the Iowa caucuses, and in 2020 The New York Times, along with most news media organizations, will again characterize the candidate with the most state delegate equivalents as the winner (though we will be reporting the results for all four measures described here).
It’s harder for there to be a big gap between the result of second alignment and the state delegate equivalency results. But not impossible.
The main way it could happen is because of turnout. The number of state delegate equivalents per precinct is set in advance, based on the average number of Democratic voters in 2016 and 2018 general elections. If a precinct has a particularly high turnout compared with recent elections, it doesn’t get more state delegate equivalents; likewise, a low turnout will not result in fewer state delegate equivalents.
So if candidates do well in relatively low-turnout precincts, they’ll get more state delegate equivalents than their vote share. And conversely, a Democrat who fares well in higher-turnout areas would tend to do worse in the state delegate equivalency vote.
The Times/Siena poll projects a caucus vote far younger, better educated and more urban than the typical Democratic vote for statewide office. If that younger tilt holds on caucus night, a candidate who excels among young and well-educated areas would probably be at a disadvantage in translating his or her votes to state delegate equivalents.
In our poll, Pete Buttigieg stands to benefit. He excelled in rural areas in our poll and widened his lead there on second alignment. This pays off even more in state delegate equivalents because these less educated areas tend to vote at higher rates in general elections than caucuses and therefore have more state delegate equivalents.
In this scenario, Mr. Sanders does well enough in rural areas, but his strength among young voters — who represent a far larger share of the caucus electorate in our poll than the Democratic voters in an average of the 2016 and 2018 general elections — costs him in the state delegate equivalency results. In some college campus precincts in this scenario, Mr. Sanders’s voters are worth only one-fourth as much as they would be in an older, more rural precinct. (It is quite possible that this cost Mr. Sanders a win in the Iowa caucuses in 2016.)
This would make Mr. Buttigieg the declared winner of the Iowa caucuses, though Mr. Sanders and Mr. Biden would probably have something to say about that.
Pledged delegates
Finally, there are pledged delegates to the Democratic National Convention, awarded based on the number of state delegate equivalents won by each candidate statewide and by congressional district.
In the past, these results weren’t available on election night either. Before this year, the state convention could elect whomever it wanted to the national convention, regardless of the results at the precinct caucus. Now, the pledged delegates to the national convention are bound to the state delegate equivalency results, so we will know right away who wins the most pledged delegates as well.
Strangely enough, the candidate with the most state delegate equivalents is not guaranteed to have the most pledged delegates.
There are two major reasons there could be a split at this late stage. One is the 15 percent threshold to win delegates in a congressional district. If the second-place candidate managed to clear 15 percent in all four districts, while the winning candidate fell short, the second-place candidate would stand a good chance of leaping ahead.
The winning statewide candidate will probably clear 15 percent everywhere, but this scenario is conceivable in such a closely fought contest and with so many candidates in the race. It could well center on the Fourth Congressional District, a mostly working-class, conservative and rural area in the northwestern part of the state. A candidate who won Iowa by dominating liberal enclaves could conceivably do quite poorly here. Nonetheless, in this simulation all of the leading candidates clear the 15 percent threshold.
The more likely way the pledged delegates could differ from delegate equivalents hinges on the number of noncompetitive candidates earning 15 percent of the vote. A candidate’s delegate haul in a district could shrink quickly if too many candidates breach 15 percent and become entitled to delegates. In our scenario, Ms. Warren breaches 15 percent in the Third District, and earns a delegate that would have otherwise gone to Mr. Buttigieg.
Another major reason is rounding. Basically, the number of delegates awarded by district or statewide is quite small. Whether a candidate earns an extra delegate here or there comes down to clearing thresholds that vary depending on the number of delegates per district.
Here’s an example. One candidate wins 51 percent of the vote in a district worth one delegate, while another candidate wins 60 percent of the vote in a second district worth four delegates. The decisive winner of the larger second district would easily win statewide but win fewer net delegates, since the second district would be a two-to-two split while the winner of the first, smaller district would win the one delegate. Winning the third delegate in that four-delegate district would require more than 62.5 percent of the vote — the lowest number that rounds up to three-fourths, rather than half.
In this case, Mr. Sanders reclaims his initial edge. Despite a big win in the Fourth District, Mr. Buttigieg can’t scratch out a delegate lead over Mr. Sanders, since only five are awarded. Mr. Sanders, on the other hand, narrowly clears a coveted rounding threshold in the First District.
In this (fairly unlikely) scenario, Mr. Sanders’s supporters would undoubtedly be frustrated. Mr. Sanders would win the most votes and the most pledged delegates, yet would be considered to have lost the state because he wouldn’t have the most state delegate equivalents. Making matters all the more frustrating, most news organizations would justify their use of the state delegate equivalency results because they award the pledged delegates, which Mr. Sanders would wind up winning anyway.
This kind of outcome would almost certainly muddle the post-election coverage and diminish the importance of winning Iowa.