We’re back with that quadrennial feature — delegate math. As we were reminded by the 2016 election, the United States uses an indirect system for electing president. Under this system, it’s not the total popular vote that counts. It’s how that popular vote translates into electoral votes. Likewise, for the nomination, the task is to turn popular votes into delegates. Back in 2008, the difference between Secretary Clinton and President Obama was the Obama’s campaign success at figuring out where they could gain an extra delegate here and there.
In a short period of time, voting will start with the Iowa caucuses. While the number of caucuses have dwindled to almost nothing (Iowa, Nevada, and Wyoming being the only real caucus states left), a caucus is different from a normal primary. In a caucus, rather than showing up, casting a ballot, and then leaving, voters actually have to take part in a meeting in which voting takes place in the middle of the meeting. The other key feature of a caucus is that voting is public, not secret. The participants literally go to different parts of the meeting facility depending upon which candidate they are supporting. Your friends and neighbors get to see who is heading to the Biden corner/room and who is heading to the Gabbard corner/room. In addition, for the purpose of any later “recount,” participants sign a pledge of support form for their candidate.
For all the states and territories, the first key number is 15%. Whether in the precinct caucus itself or at the congressional district level or at the state level, a candidate needs to get 15% of the vote to win delegates. (Some smaller precincts have higher thresholds because they are only selecting two or three delegates, but 15% is still a good general rule of thumb in looking at polling numbers.) What makes a caucus different than most primaries is that there will be a chance after the initial division of the precinct into the separate candidate areas for voters to realign.
In Iowa, the meetings on February 3 are precinct caucuses –meetings at the sub-county level. In addition to the February 3 meeting, there are “satellite” caucuses for those — both in-state and out-of-state — who are unable to be present on February 3 at their precinct caucus. Unlike Nevada (where the early participants cast votes that count toward their local caucus), these “satellite caucuses” are separate votes that result in the selection of its own group of congressional district and state convention delegates. (Generally, the satellite caucuses are generally supposed to take place at the time of the precinct caucus, but the state party can approve meetings at different times on February 3.)
One of the key facts of delegate math in Iowa is that the ultimate division of state and congressional districts is based on the number of “state equivalent” delegates that a campaign wins in the precinct and satellite caucuses. While the party will report the raw vote totals — both the first expression of preferences and the final expression of preferences — the number that counts is the state delegate equivalents.
The calculation of state delegate equivalents has three parts. The first part is the actual precinct caucuses. Each county within each congressional district has a number of state (and congressional district) delegates based on the most recent election for governor and the most recent election for president — how many Democratic votes were cast in that county in those two races. Each county then uses a similar formula to determine how many delegates each precinct gets to the county convention. Because of rounding rules (and the fact that even the smallest, most Republican precinct gets at least one county convention delegates), there are slight differences in the weight that county conventin delegates carry in different counties. (In other words, in one county, you could have ten county convention delegates for each state convention delegate, but in another county you might have eleven county convention delegates for each state convention delegate.)
Second, for in-state satellite caucuses, each satellite caucus is treated as a precinct within a fictional satellite caucus county within the congressional district. Depending on the total number of participants in the satellite caucuses within that district, the satellite caucus county receives extra district convention/state convention delegates (expressed as a percentage of that district’s current number of district/state delegates). In turn, based on the number of participants in each satellite caucus, the satellite caucus receives a certain weight of (fictional) delegates to the (fictional) satellite caucus county convention. Again because of rounding rules, four county convention delegates in one congressional district will probably not be equivalent to four county convention delegates in a different congressional district.
Finally, out-of-state satellite caucuses are treated as an “at-large” county. (Apparently, the state convention delegates allocated to this “county” will only count toward the state-level allocation of delegates.) Depending on the total number of participants in out-of-state caucuses, this at-large county will receive a certain number of state convention delegates (somewhere between one and six). Additionally, each individual out-of-state satellite caucus will receive a certain number of fictional “at-large” county delegates based on the attendance at the individual caucus. Because there is a maximum and a minimum number of “county” delegates that each individual caucus and a minimum and maximum number of state delegates that the out-of-state caucuses as a group receive, the state-delegate-equivalent weight of each county delegate in “at large” county may be different than the state-delegate-equivalent weight in other counties.
In short, while each precinct caucus and satellite caucus will allocate county convention delegates, not all county convention delegates are equal. So calculating the results of state-delegate equivalents is not as simple as counting the total number of county convention delegates won.
This year will be difference than past cycles in that the party will also be reporting raw numbers. The raw numbers will be different than state delegate equivalents in several ways. First, the initial preference vote (the vote that would count in most states) will probably have two groupings of candidates. One grouping will feature candidates in excess of fifteen percent; the other grouping will feature candidates with less than ten percent. Because it takes fifteen percent in a given precinct to win county convention delegates, the second preference vote (the vote after realignment) will probably have very different results with candidates who failed to crack fifteen percent in the initial vote dropping to close to zero and the candidates with more than fifteen percent in the initial vote gaining votes. (It will be interesting to see the exact shift.)
Second, even with the realigned vote, the delegate counts will be somewhat different. The difference comes from two features of the process. On the one hand, the delegate selection process involves rounding. In a precinct that elects ten county convention delegates, a candidate will get two delegates if they get 22% of the vote in that precinct or if they get 17% of the vote in the precinct. Over the state as a whole, one would expect the rounding to roughly even out, but there is no guarantee. On the other hand, delegate math does not care about the turnout in any individual precinct (other than the satellite caucuses). County convention delegates are based on votes in past election. So a precinct has the same number of county convention delegates if 10% of Democrats turn out on February 3 or if 40% of Democrats turn out on February 3. So, for example, high turnout in Sanders-leaning precincts will not alter the state-delegate equivalents won but would increase the raw vote totals for Senator Sanders.
The big difference in Iowa this year is that — as mandated by the national rules — national convention delegates will actually be allocated based on the results of the precinct caucuses. In past cycles, the state delegate equivalents were a useful rough tool for estimating national convention delegates. But, the actual awarding of delegates took place later. But, that rough estimate had to be run through the county conventions — where delegates from low performing candidates could switch side and the leading candidates might not get all of their delegates to attend — to get to actual state and district convention delegates. And even with actual delegate numbers (as opposed to delegate equivalents), things could still change at the state and district conventions due — again — to delegates having to realign from non-viable candidates and some delegates failing to attend. This cycle, there will be no post-caucus shifts unless candidates drop out. (If a candidate does drop out, the numbers are recalculated for the purposes of allocating at-large and pledged party-leader delegates).
As in other states, Iowa’s national convention delegation consists of four groups of delegates — automatic delegates (members of Congress and DNC members), district-level delegates, pledged party leader delegates, and at-large delegates. A new feature is that automatic delegates will not have a vote on the first ballot at the national convention, but these delegates can opt to run for one of the pledged positions. If an automatic delegates wins a pledged delegate position, the delegation effectively loses one delegate but the automatic delegate gains the right to vote on the first ballot. (This change could be a “be careful what you wish for” change to the process. One of the reasons for automatic delegates was to free up slots for ordinary grass roots activists to attend the convention without having to compete for delegate slots against party leaders.)
Before going into the allocation of delegates, there is one necessary caution. Because Iowa is first, there will probably be a higher number of viable candidates than will occur in later rounds. With two candidates, rounding fractional delegates is easy. If candidate A wins 4.6 delegates to candidate B’s 3.4 delegates, the fractional delegate goes to candidate A (or vice versa if the numbers were 4.4 and 3.6). With multiple candidates, you could have several fractional delegates (rather than one) and using .5 is the threshold could award too many or to few delegates (e.g. with ten delegates available and the candidates winning 4.7, 3.65, and 1.55 delegates rounding them all up gives you eleven delegates. bit with the candidates winning 4.4, 3.32, and 2.28, rounding down only gives you nine delegates). Additionally, there can be districts in which the threshold for winning delegates (15%) is lower than the quota (i.e. the number that gives a candidate a “whole” delegate — basically 100%/number of delegates available). That creates the possibility of more candidates winning delegates than there are delegates available, but it also creates the possibility that a candidate could receive less delegates than the number of “quotas” they have won.
For the congressional district level, Iowa has four districts. Iowa’s Third District has eight delegates. With a minimum of 15% to win delegates, any candidate who reaches that threshold will exceed the “quota” for a delegate. With eight delegates, the quota for a delegate is 12.5% . Thus, 15% wins at least one delegate, 25% wins at least two delegates, 37.5% wins at least three delegates, 50% wins at least four delegates, 62.5% wins at least five delegates, 75% wins at least six delegates, and anything over 85% wins all eight delegates.
Iowa’s First and Second Districts have seven delegates each. Again, that places the per delegate “quota” below 15% (roughly 14.3%). Thus 15% wins at least one delegate, 28.6% wins at least two delegates, 42.9% wins at least three delegates, 57.2% wins at least four delegates, 71.5% wins at least five delegates, and anything over 85% wins all seven delegates.
Iowa’s Fourth District (a/k/a Representative Steve King’s district) has five delegates. As such, the quota for this district (20%) is above the threshold for viability. However, it is unlikely that there will be any significant complications due to this unless six candidates reach 15%. Similarly, Iowa has five “pledged party leader” delegates. For both the Fourth District vote and for the state-wide votes for the PLEO delegates, barring six candidates getting 15%, 15% wins at least one delegate, 40% wins at least two delegates, 60% wins at least three delegates, 80% wins at least four delegates, and 85% wins all five delegates.
Finally, Iowa has nine at-large delegates. Again, the 15% threshold exceeds the roughly 11.1% quota for delegates. Thus 15% state-wide will win at least one delegate, 22.2% wins at least two delegates, 33.3% wins at least three delegates, 44.4% wins at least four delegates, 55.5% wins at least five delegates, 66.6% wins at least six delegates, 77.7% wins at least seven delegates, and 85% wins all nine delegates.
Of course, with this many candidates running, it is unlikely that any candidate will sweep all of the delegates in any of the district. The real question is going to be for those candidates getting near 15% on the initial vote. It is likely that such candidates will barely make the threshold for viability in some precincts but miss viability in other precincts. And missing in any substantial number of precincts will make it difficult to get to 15% at the state-wide or district level in state delegate equivalents. And, thus you could have some candidates who would qualify for delegates on the initial vote but ultimately fail to qualify for delegates when the vote is translated into state-delegate equivalents. And, as will happen in other states that do not use rank-choice voting, calculating delegates requires adjusting the vote totals to only count the vote (here the state-delegate equivalents) for those candidates that qualify. (Of course, it is possible that some candidates might qualify on the state level but miss in one district or qualify in one district but miss on the state level. As such, there could be different adjustments for each district and for state-wide totals.)
