David P. Redlawsk: The Iowa Caucuses are Like the Electoral College
[David P. Redlawsk is Associate Professor of Political Science at the University of Iowa and Director of the University of Iowa Hawkeye Poll. ]
Talking with some folks here in Iowa the other day, we began to think about how the Iowa Caucus mirrors that other great institution of presidential politics, the Electoral College (EC). Now this is ONLY for the Democrats. The Republicans approach their presidential preference process quite simply – they vote. That is, when Republicans arrive at their caucuses, there will be some speeches and discussion about the candidates, and then they will just write their choice on a secret ballot, which will be counted, and the precinct results reported to the Republican HQ in Des Moines, where they will add up all the precincts and announce the final vote. Simple. Fast. And pretty easy to understand.
Democrats, though, are different.
Democrats are electing delegates to county conventions (to be held in March) and those delegates are elected roughly (and the key word is roughly) in proportion to the support each candidate receives at each precinct. So if Dennis Kucinich (right) gets 45% support in a lefty Iowa City precinct, he (should) get 45% of the delegates that precinct elects to the Johnson County Convention.
How, you ask, do the Democrats know how many delegates to elect at each of the 1784 precincts in Iowa? That, my reader, is determined by a mystical formula that each individual county applies (with oversight from the state party). So let me take a shot at it and explain why the Caucus is like the Electoral College (EC). Here goes…
First, like the EC where we know how many votes each state has, we know the delegate counts for every precinct before the caucuses happen – they are set ahead of time based on the Democratic performance in the precinct in the two general elections prior considered as a percentage of the total county democratic performance. (So for example, my precinct might have 3.2% of the Democratic votes in my county over the past two elections.) That percentage is then applied to the total number of delegates at the county convention to determine the delegate count for the precinct. You cannot, however, compare these numbers across counties, because each county can set its convention at whatever size it wants to. Within each county then, some precincts might elect 12 delegates and others 1 delegate. The key point is that everyone knows ahead of time how many delegates each precinct will elect, AND that number is NOT related to the number of people who actually show up to caucus, just as the EC vote is not related to the number of people who turn out in each state in a presidential election.
So to take my own county, the Peoples’ Republic of Johnson County, in some precincts in 2004 it took as few as 25 caucus attendees to win one delegate while in others it took 75 or more to win one delegate.
Why does this matter? Because the results reported by the state party on caucus night are NOT the actual votes cast for each candidate. Rather, they are “State Delegate Equivalents.” Essentially, since we DO know how many delegates each county will get to the STATE convention, and we know what percentage of the state convention each county represents, the delegates at the county level can be converted to State Delegate Equivalents by doing some basic math, which the party will do before they report the results as percentages for each candidate.
Unlike the Republicans, the Democrats have not released the actual number of people supporting each candidate in the past, so we do not get the vote count. When you get told on caucus night that Kucinich won 38% and Gravel won 18% and Clinton won 5%, it means each won the number of delegates to county conventions across the state that calculate to that percentage of state convention delegates. (Simple, isn’t it?)
Because the actual vote counts are not reported, there is no research that looks at the relationship between headcount (votes) and delegates. Can’t do it without all the data! My guess is that in fact across the whole state the numbers are reasonably related - simply because the overage/underage in terms of votes needed to earn a delegate may wash out across the whole state. BUT, that’s not completely true if one candidate (say Howard Dean last time) has disproportionate support in precincts where voters do NOT turn out regularly in the general election (so Democratic performance is lower and the number of delegates is thus lower) but a whole lot of folks suddenly show up to caucus. I’m thinking here of student precincts in 2004. My guess (and it is only a guess) is that Dean got a lower percentage of delegates than he would have in a system that just reported the vote count. Likewise, if a candidate has a strong rural strategy where there may only be one or two delegates per precinct but it takes only a handful to people to win them, he or she can get a larger percentage of delegates than actual statewide votes
And of course, candidates can get support but no delegates. The existence of the 15% viability threshold (and even higher in small precincts that only elect two or three delegates) guarantees this. In an extreme example, a candidate could get 14.9% support in every one of the 1784 precincts this time, and would be reported with 0 delegates, and thus 0%.
So how is all this like the EC? Well, the EC votes do not depend on the number of people who show up to vote; the geography of the election matters – just look at the red/blue maps from 2000 and 2004 to see that; and the EC vote generally overstates the actual vote of the winner and understates that of the loser. Maybe it is a bit of a stretch, but in the end, the point is to try to understand what Iowa is telling us on the night of January 3. If you’re a Republican the results are clear. If you’re a Democrat, well…
Disclosure: I’m a Democrat, an activist one. I was the acting County Chair for Johnson County during the 2004 caucuses and was thus responsible for 57 precinct caucuses that year. How do I defend what the Democrats do? Well, it’s pretty simple. We are electing folks to a convention, not voting on presidential candidates. Because the media wants to know “who won,” we report who won – but since it is the conventions that actually will ultimately choose delegates to the National Convention, which will (in theory of course) be the decider about the nominee, it is more accurate to report who won how many delegates than it is to report who got how many votes.
Read entire article at Britannica Blog
Talking with some folks here in Iowa the other day, we began to think about how the Iowa Caucus mirrors that other great institution of presidential politics, the Electoral College (EC). Now this is ONLY for the Democrats. The Republicans approach their presidential preference process quite simply – they vote. That is, when Republicans arrive at their caucuses, there will be some speeches and discussion about the candidates, and then they will just write their choice on a secret ballot, which will be counted, and the precinct results reported to the Republican HQ in Des Moines, where they will add up all the precincts and announce the final vote. Simple. Fast. And pretty easy to understand.
Democrats, though, are different.
Democrats are electing delegates to county conventions (to be held in March) and those delegates are elected roughly (and the key word is roughly) in proportion to the support each candidate receives at each precinct. So if Dennis Kucinich (right) gets 45% support in a lefty Iowa City precinct, he (should) get 45% of the delegates that precinct elects to the Johnson County Convention.
How, you ask, do the Democrats know how many delegates to elect at each of the 1784 precincts in Iowa? That, my reader, is determined by a mystical formula that each individual county applies (with oversight from the state party). So let me take a shot at it and explain why the Caucus is like the Electoral College (EC). Here goes…
First, like the EC where we know how many votes each state has, we know the delegate counts for every precinct before the caucuses happen – they are set ahead of time based on the Democratic performance in the precinct in the two general elections prior considered as a percentage of the total county democratic performance. (So for example, my precinct might have 3.2% of the Democratic votes in my county over the past two elections.) That percentage is then applied to the total number of delegates at the county convention to determine the delegate count for the precinct. You cannot, however, compare these numbers across counties, because each county can set its convention at whatever size it wants to. Within each county then, some precincts might elect 12 delegates and others 1 delegate. The key point is that everyone knows ahead of time how many delegates each precinct will elect, AND that number is NOT related to the number of people who actually show up to caucus, just as the EC vote is not related to the number of people who turn out in each state in a presidential election.
So to take my own county, the Peoples’ Republic of Johnson County, in some precincts in 2004 it took as few as 25 caucus attendees to win one delegate while in others it took 75 or more to win one delegate.
Why does this matter? Because the results reported by the state party on caucus night are NOT the actual votes cast for each candidate. Rather, they are “State Delegate Equivalents.” Essentially, since we DO know how many delegates each county will get to the STATE convention, and we know what percentage of the state convention each county represents, the delegates at the county level can be converted to State Delegate Equivalents by doing some basic math, which the party will do before they report the results as percentages for each candidate.
Unlike the Republicans, the Democrats have not released the actual number of people supporting each candidate in the past, so we do not get the vote count. When you get told on caucus night that Kucinich won 38% and Gravel won 18% and Clinton won 5%, it means each won the number of delegates to county conventions across the state that calculate to that percentage of state convention delegates. (Simple, isn’t it?)
Because the actual vote counts are not reported, there is no research that looks at the relationship between headcount (votes) and delegates. Can’t do it without all the data! My guess is that in fact across the whole state the numbers are reasonably related - simply because the overage/underage in terms of votes needed to earn a delegate may wash out across the whole state. BUT, that’s not completely true if one candidate (say Howard Dean last time) has disproportionate support in precincts where voters do NOT turn out regularly in the general election (so Democratic performance is lower and the number of delegates is thus lower) but a whole lot of folks suddenly show up to caucus. I’m thinking here of student precincts in 2004. My guess (and it is only a guess) is that Dean got a lower percentage of delegates than he would have in a system that just reported the vote count. Likewise, if a candidate has a strong rural strategy where there may only be one or two delegates per precinct but it takes only a handful to people to win them, he or she can get a larger percentage of delegates than actual statewide votes
And of course, candidates can get support but no delegates. The existence of the 15% viability threshold (and even higher in small precincts that only elect two or three delegates) guarantees this. In an extreme example, a candidate could get 14.9% support in every one of the 1784 precincts this time, and would be reported with 0 delegates, and thus 0%.
So how is all this like the EC? Well, the EC votes do not depend on the number of people who show up to vote; the geography of the election matters – just look at the red/blue maps from 2000 and 2004 to see that; and the EC vote generally overstates the actual vote of the winner and understates that of the loser. Maybe it is a bit of a stretch, but in the end, the point is to try to understand what Iowa is telling us on the night of January 3. If you’re a Republican the results are clear. If you’re a Democrat, well…
Disclosure: I’m a Democrat, an activist one. I was the acting County Chair for Johnson County during the 2004 caucuses and was thus responsible for 57 precinct caucuses that year. How do I defend what the Democrats do? Well, it’s pretty simple. We are electing folks to a convention, not voting on presidential candidates. Because the media wants to know “who won,” we report who won – but since it is the conventions that actually will ultimately choose delegates to the National Convention, which will (in theory of course) be the decider about the nominee, it is more accurate to report who won how many delegates than it is to report who got how many votes.