Conventions Post-Mortem: It's the Turnout, Stupid!
(I originally wrote this article for the online blog The American Conversation. It was posted on September 10th, 2012, right after the Democratic Convention was over. In it, I described the Median Voter Theory, and how that presented the Democrats with a favorable election climate in 2012. I also predicted that because of that favorable structure, the Dems would win the White House in 2012. This article has, to this point, been archived at Ko Point, though this version also includes the important graphic which is missing in the archived version. My thanks to Dan Patterson for holding onto it.)
The Conventions are over. But you may be scratching your head wondering what happened. Well, if you feel more or less likely to go and vote after the convention, then they got you. Here's how.
We should start off this discussion by dealing with the sticky subject of the "Independent" or the "Undecided" voter. Lots of people like to call themselves that, but do they have as big an influence on the outcome of a political campaign as people are led to believe?
If you really aren't interested in the question, let me answer right off the bat. No. They don't. It's not because the opinions of a sizable portion of the electorate won't affect the outcome. Work done on the question has recently been popularized by NPR, on both Morning Edition and Talk of the Nation. These segments more or less reiterate what some political scientists have tried to point out for a long time: political independents don't matter much because they don't really exist. If you think of yourself as an "independent," you actually aren't.
As a partisan of a certain stripe myself, I get it. There is a vogue to calling one's self "independent". Being "independent" means that no Party controls you. Or it means you get to defy the label people will attach to you when they discover what your Party affiliation is. Most of the people who call themselves "independent" don't think it is a calculation. Most people who call themselves "independent" really do think they are independent, that they are free of the two major Parties, and will make a decision about who to vote for based solely on what each candidate believes, and not the Party that this person is associated with.
Problem is: they don't actually do that. They are as nonobjective in their vote as the person who calls themselves a Democrat or a Republican. They have made their minds up who they will vote for long before they have heard the case for and against that individual. They have an idea in their heads about what a good society looks like and they vote for the guy or woman who they believe will get them closest to that vision. In other words, they are just like the rest of us. They are no more special or intelligent or objective. They are just the same.
And the same goes for the so-called "undecided voter." That person does not really exist unless they have been hiding under a rock for a year. Modern political campaigns are designed to eliminate indecision in voters. I submit to you--and this would be a bet I would be willing to take if there was some way to verify it-- that if you are undecided at this point in the campaign, you will either not vote, or will vote randomly in two months. So there is really nothing the candidates can say or do to ensure that they will get your vote. You will give it to them or their opponent based on some entirely non-political calculus that is known to you and you alone.
A SIDE NOTE ON "THE HORSE RACE"
As a slight deviation from the point of this article, the fact that people still insist that they are either independent or "undecided" when they really aren't causes problems for vote counters in both campaigns. If I am counting votes to try to come up with an electoral strategy, I can, with almost scientific accuracy, predict the votes of those who declare that they are either "Democratic" or "Republican." Luckily, most of them do not lie when asked who they will be voting for.
But vote counters have no reason to trust that "independents" or "undecided" voters are telling the truth, because in picking their label, most of them aren't being entirely honest with the people asking them the questions. Consequently, there is no reason to assume that someone who will tell them that they are independent or undecided are actually telling them the truth. And since it is apparently fashionable to at least appear to be thoughtful and considerate and ambivalent rather than decisive in this society, there are millions and millions of people who will claim that they are one thing and are actually another thing. It is therefore very difficult to get an accurate count--the safe assumption is that a third of the electorate is lying at any given time when asked who they plan on voting for.
Incidentally, this is also why the campaigns look like a horse race, with one candidate up one week and another candidate another week. As the campaign goes on, as if by magic, the turbulence amazingly dies down and, just before the elections, a clear front runner emerges. It is not that the campaigns are improving their standing among the population during the campaign. We will find out below that there is little a candidate can actually do to convince voters to change their minds about who they will vote for. It is instead that 1) those who call themselves independents and undecideds are not accurately reporting their plans to pollsters until all the evidence appears to be in to justify their secretly held preference, making indecision at last an untenable position, and 2) Each pollster gets a different mix of these people every time the question is asked. Occasionally, some unforeseen event changes the course of an election, but for the most part, elections are decided once the contestants are decided, and we just don't know it.
Campaigns benefit from this observed but ephemeral turbulence: they get to front like they are gaining more and more support and present the appearance of momentum, but more important than this appearance is it gives them time to convince people to come out and vote or not, as I discuss below. The Media benefits from this as well: it gives them something more exciting to report on than a massive "Get out the vote" effort on the part of political candidates. And the people benefit from it because it makes the practice of electing people to make decisions for you a bit more exciting.
Win-win win, unless you are a campaign manager trying to craft an electoral strategy, and your efforts are constantly frustrated by people who simply refuse to honestly report who they are going to vote for because they think they are smart, for some reason, to keep their cards close to their chest. We could have short campaigns run relatively inexpensively, and therefore ones which are far more accessible to the average citizen if we knew early how the election was going to shake out, and we would know that if perhaps as many as 75 million voters in a given year would simply just report honestly about what they intended to do, rather than waiting until the last possible minute to declare. I return now to the subject at hand.
So, since we know that independents and undecided's don't actually exist, what the heck were these last two weeks about? I argue that they were about turn out. Before I get too far into this, let me post up a chart that I created to illustrate the ideas I will lay out below. The theory is not entirely mine. It is based on median voter theory, the platform convergence ideas of Hotelling, the groundbreaking economic theory of Anthony Downs, and the great 2004 work on American voter behavior by James A. Stimson, called Tides of Consent. It is probably more appropriate to say that I am popularizing work done by others, as my expertise is in foreign policy, not something like voter behavior.
Median voter theory states that the Party which captures the median voter--that is, the point on the political spectrum where exactly half the voters are to the ideological left of that voter and half of them are to the ideological right in a given election-- that Party will win the election. If they can craft a message which will capture voters significantly to the other side of that median from their own base, without losing too many of their own base, they will have a landslide victory. If, on the other hand, the opposing Party is able to capture the median voter and a significant number of your traditional voters, you will lose.
Moving along, here is the chart which describes voter behavior, both in a generic election and in something which looks more like the 2012 election.
This chart depicts several things. The x-axis (horizontal) is some ideological position of any given voter. The system has some number of total voters who will vote for both of the major Parties, each of them has a unique set of ideological preferences, on a left-right spectrum. Along that axis there are several marks. The blue vertical line depicts the ideological preferences of the median Democratic voter, that is, half of those people who could support a Democratic candidate are ideologically to the left of that position and the other half are ideologically to the right of that voter. The GOP has a corresponding line. That line can be called the "position of the Party." There is also two other median voters marked: the ideological position of the median voter in an "Issues" election and the ideological position of the median voter in a "Symbols" election. Finally, the line between those two points is literally the person who is dead center, ideologically, of the electorate.
The y-axis is a range of chance, from 0% at the x-axis to 100% on the "1" mark, that any given voter will vote for a given Party. The parabolic lines are the percentage where a given point on the x-axis corresponds with a given percentage chance that the person will vote for that respective Party. The Blue parabola is the chance line for the Dem Party, and the Red parabola is the chance line for the GOP. Predictably, there is something like a zero percent chance that a person to the left of the Dem Party's median ideology, its Party's position, will vote for a Republican, and vice versa. But voters between those two ideological positions have a larger chance of voting for one Party and a correspondingly small chance of voting for the opposite Party. One can guess that the closer one is to the Party's median ideological position is more likely to vote for that Party and less likely to vote for the other Party.
A true Independent, the person who is equally likely to vote for either Party is the person who is ideologically equidistant from both Party's positions. That is only one person-- Who that is, nobody will know, but there was a fictional movie about it starring Kevin Costner in 2008, like it would be that easy to determine who that would be. He is the Median Voter of the theory, and without that stuff about Issues and Symbols discussed about, the Party which captured him is the victor.
It is important to note that there are a number of assumptions built into this model. First, you will note that it is very sparse on numbers. This is because the model is meant to illustrate a concept, rather than serve as a precise prediction tool. I would have to have a LOT more data to develop a graph which accurately depicts the US electorate in 2012. This merely offers a graphic depiction of an idea, but it is useful then in that it allows one to see how it could work in a number of generic examples, instead of tying us down to one election and forcing us to go find a mountain of data for every election. Secondly, it assumes that voters vote for their preferences. In other words, it assumes that voters are rational. But people do things for a lot of different reasons, and in some cases, choice is made randomly. There is no accounting for this sort of behavior, but if we assume that voters actually go in and vote for the things they want, a society that has that feature would look something like the models I presented above. Finally, this model assumes things about the political system which I don't really need to go into here. For example, there is no electoral college built into this system. There is also nothing like weather delays or ID card laws or so forth which may lower the chances that I would be willing to vote for one candidate rather than another.
Limitations of the model aside, this model does offer some explanation for voter behavior which is more empirically supportable than some other models, and it is also helpful to simply illustrate the concept and the theoretical explanation. There are theories which are more useful, but for our purposes, this model seems to describe political choice in a [ed. 2 Party, representative] democracy fairly well.
"ISSUES" ELECTIONS VS. "SYMBOLS" ELECTIONS
This distinction between "Issues" and "Symbols" elections deserves a little discussion. The American electorate has a bit of a split personality : If you ask them whether or not they support what we call "liberal" policies, like social spending, medicare, poor relief, and so forth, they usually answer "yes." Liberals tend to do very [well] when they are able to explain policy to people, because people generally see the need for those policies.
However, during some year, symbols are more important to the electorate. For some reason, if the election is about things like "America" or "freedom" or the Flag, or if a candidate is associated with an eagle or something like this, people respond favorably to that as well. This is why conservatives had such luck in 2002 and 2004 elections. Those elections were not about policies, but about symbols traditionally associated with "America," especially the war on terrorism, 9/11, and national security. If the election is about issues, Democrats tend to do better. If it is about symbols, Republicans tend to do better. To the right of the median for a "Symbols" election, those voters are reliably Republican, to the left of the median for an "Issues" election one will reliably find Democratic Voters. The so-called "Independents" are between those two points, but as we will see, Party ideology throws those voters into one or the other camp whether they like it or not.
For each election, either issues or symbols are more salient for the electorate, and so this factor forms a major structural influence on the outcome of the election: the astute political campaign will understand this salience, and take appropriate action in response, playing up their strengths or attempting to compensate when they are in a weak position.
HOW THE MODEL PREDICTS VOTE TOTALS
The model works by finding the ideology of a given voter and using that to figure the likelihood of them voting for one Party or another. In the top chart, a perfectly divided society is depicted, where the median voter is actually the center of the ideological distribution of both major Parties. The Party which is successful in capturing the vote of this person will win the election, for a number of reasons which I will not get into, except to say that to win, a candidate needs 50% +1, the +1 being the median voter. In that chart, the Parties are both equidistant, policy-wise from the center of the electorate, and both have an equal chance of capturing the median voter, as well as a similar, mirrored chance of capturing the other side's respective voters as they do themselves.
This first chart represents at once a campaign manager's dream and democracy's nightmare. The campaign manager has to only craft the message to suit one person, but 50% -1 of the electorate are entirely cut out of the election, while one person decides everything. Luckily for us all, but not for campaign managers, elections never turn out this way. In issues elections, that median voter shifts significantly left, and has a much higher chance of voting for Democrats than they do Republicans. The opposite is true in symbols elections. Consequently, campaigns are tailored to a much broader swath of people, and more people matter with regards to the outcome.
Movement left or right of the median ideological position as issues and symbols respectively become more or less salient to the population does interesting things for the Parties themselves. It allows Parties to move their platforms to the left or right too- potentially giving them coverage of more ideological positions and therefore increasing the chances that any one given voter, including the median voter wherever that person happens to be on the spectrum, will vote for their Party. If symbols are more salient, for example, this allows the GOP to move its own median position to the right, because it now has a much higher chance of capturing the median voter and so has a bit of freedom to pander to its right. Meanwhile, the Democratic Party has a lower chance to capture the median voter, so it will also shift to the right as well to try to capture some of the middle which is left behind after a rightward shift of the GOP.
But a shift toward the middle by the Democratic Party will also decrease the potential that a given leftist, for example, who would never vote for a Republican anyway, will vote for the Democratic Party-- a number of other options now become available for those people, like voting for a third Party or not voting. The more the Democratic Party moves right, the more of the middle they capture, but the more of their base they lose. Conventional reason says that there are more voters in the middle than in the base, so some losses in the base can be offset by small gains in the middle, while losses in the middle are difficult to make up in the base. Apparently, the GOP disagrees, as we see below.
Over the first decade of the 21st century, symbols were quite important. Policy was dominated by symbols: terrorism, national security, the Flag, everything. Whether or not Obama wore a flag lapel pin was a major controversy in early 2008. During most of this time, the GOP, comfortable in the structure which suited it electorally, safely moved their platform to the right, and the Democrats were initially slow to respond. The GOP continued to move it to the right through the entire decade, greatly opening up the middle to be exploited first by the Congressional Democrats in 2006 and then by Obama in 2008 (to the chagrin of most progressives).
But then something happened: the 2008 election turned into an issues election as the housing crisis exploded, the banks and car factories went under, and the jobs market collapsed. People were ready to get out of Iraq, and the symbols associated with that and the war on Terror were no longer as salient for voters. The structure was changing. The GOP suddenly was out of step with the median voter, and the Democratic Party, to their credit, finally read the writing on the wall. They were well positioned close to the median "issues" voter in early 2008, crafted a centrist platform which exploited that open center vacated by an immoderate Republican Party, and Obama easily won the election.
The rise of the TEA Party from the GOP's right in early 2009 was [ed. in a small part, anyway] a reaction to the structural change that happened in 2008. Since the median voter moved toward the "issues" position and away from the "symbols" position, the Democratic Party could afford to move left, and in a normal year the GOP would follow them. But the pressure the TEA Party put on the GOP in the meanwhile prevented that move. The Dem Party moved a little to the left, as they could be expected to, while the GOP actually moved to the right.
Consequently, there was a smaller chance that either Party would claim the median voter, which was somewhere between the dead center of the electorate and the "Issues" position. The turnout numbers in the 2010 midterm election demonstrate this reality. The reason that the GOP won in 2010 was because the TEA Party was able to make the election a referendum on Obama's first two years, but also, the Democrats failed to turn out centrist voters. The GOP did not win that election then, though they swept into the House. The Dems lost it. We could say that the GOP did less bad than the Democrats in 2010.
THE 2012 ELECTION- DECIDED BEFOREHAND?
Following the disaster for Democratic Legislators in 2010, the GOP and the Dems should have moved closer together: the Dems because they sought to prevent a repeat in 2012 and therefore, need to move right to increase the chance of picking up that median voter, and the GOP because they were out of step with the structure of the electorate, but also because they had a wider group of people that they had to cater to given their increases in representation. They had a wider electorate which includes more people on the "left" than who voted for them, but no clear mandate. The Dems did what they were theoretically expected to over the next two years, and upset a lot of their base in the process (this is thought to be a problem in 2012- they moved so far to the center that people who thought they would move to the left and stay there are disappointed.)
But the GOP did not. They not only failed to move to the center, but the TEA Party and the supporters of austerity pulled them further to the right! And this election is not about symbols- it is all about issues, and both Parties agree. You can do the math. The chance that the GOP will capture the median voter is incredibly thin, because their positions are so far to the right that the Democratic Party has a HUGE range that they can play with and still guarantee that they have way more than a 50/50 chance of getting not only the median "issues" voter, but also the center of the political electorate AND, if they go right far enough, [ed. a small chance of winning] the median "symbols" voter too, as my chart suggests!
For the GOP, this offers an early explanation on the so-called "tack to the center" of the Romney campaign that is appearing in some sources, especially his Meet the Press Interview from Sep. 9. Unfortunately, for him, it is probably a bit too late, because any move to the center squanders the excitement among the base which was generated by his choice of running mate [ed. the bonafide conservative Paul Ryan. Such a tack to the center] will not change anyone's mind, but will dampen turnout. But it was almost as if someone in his campaign finally persuaded his managers to look at median voter theory. You can imagine that if they understood this earlier, they wouldn't have gone so far to the right in the first place. [ed. But given attitudes of Republicans in 2012, then Romney might not have gotten the nomination in the first place.]
For the Dems, this election is an ideological balancing act. Do they go farther right and catch more of the center, and in the process, decrease the likelihood that their base will vote for them, or do they adopt a middle policy, and keep as many voters on the left and right as they possibly can? It seems that they are taking the second course, not going as far as they could (as depicted in the second graph) to the right, but certainly exploiting the failure of the GOP to understand the structure of the electorate in 2012.
THE CONVENTIONS- BOOSTING TURNOUT
Which brings us back to the function of the Conventions. This post initially was about the conventions, and I got into theory. But you will see, it is a profitable diversion, because it explains, as theory, what these people were doing at the convention. We know a few things. We know that independents don't exist in any real sense.
We know that undecideds don't really exist either. We should understand, therefore that the conventions were not designed to convince anyone to vote for Romney or Obama, which is something they would need to do if there was a large group of people who still actually haven't decided.
But luckily, whether or not a person will vote for one candidate or another is largely determined by his ideology in relation to the Party's platform and the structure of the election. Each one of us has a given chance that we will vote for Obama or Romney based on these factors which has nothing to do with how much money they spend on television ads. We would still have that percentage chance of voting for them whether they campaigned a single day or not.
But surely then, there has to be some other reason that these people are putting so much time and effort into campaigning and having conventions. My educated guess is that the reason why they do all this stuff is because this year, as it is in every year, it is about turnout. Given that the electorate has a set of characteristics which exist regardless of if a single dollar is spent on the campaign, all of that money and all of those speeches are designed to increase the percentage chance that you will turn out to vote. Unfortunately, I do not have a commonly used model to determine the percentage chance a given person will turn out to vote. There are a number of things which are quite beyond our control: the laws of the polling place, the weather, whether or not your kid is sick. We are not interested in those- as they are quite beyond the control of political campaigns.
So they put their money and their effort into the things they can control. Their job is to let you know that there is an election coming up, and the guy you are predisposed to vote for anyway is worth the time and effort you will take to go and vote for him. The speeches at the GOP convention and the Dem Convention were "good" or "bad" to the degree that they accomplished those two goals. All commercials you will see from here on out are good or bad to the degree that they do those two things.
The first one is easy. Election Day is listed on almost all of our calendars. But candidates will constantly remind you that you need to go vote at this particular time. And there is a massive Get Out the Vote campaign associated with each of these guys to ensure that you follow through.
The more difficult task is to ensure you that you are making the right decision when you act on the percentage that you will vote for Candidate A over Candidate B. Romney chooses the tactic of making Obama look like a worse than the alternative. If you have a 5% chance of voting for Obama anyway, a dire prediction of a dystopic second Obama Presidency may be enough to get you to the polls. You aren't going to vote for Obama anyway; that's pretty much a foregone conclusion.
But whether or not you decide to go to the polls is not so foregone-- Romney's campaign exists to make sure you turn out and do what you are highly likely to do anyway, which is vote for him when you get to the polls. Obama takes a different approach in talking about how great he has been and will be in the future, but the goal is the same-- getting those people who are likely to vote for Obama to go and vote. Essentially the winner is the person who is able to turn out the most votes.
There is a flipside to a "Get Out the Vote" Campaign. A candidate will do himself a favor if he is able to undermine the voter turnout for the other guy. They are not going to convince anyone to change their mind: the goal must be to get your opponent's voters to stay home (or to be denied a vote which produces the same result.) Predicting a dire future under your opponent is a fairly effective way for those who are in the middle, and don't have their own compelling reasons for picking one or the other. A full harangue at a Party's convention, or a good showing at the debates won't affect the opposing candidate's base at all, but it could have an effect on turnout in the middle.
Which brings us back full circle to the beginning. This is why the observers all say that the middle matters. Turnout around the median voter DOES matter, without a doubt because they are the most likely to not have strong opinions about the candidate they will eventually vote for, and therefore are most susceptible to being encouraged to stay home. If you can get the middle to not turnout, then the base that the Party sacrificed to run a centrist campaign become that much more damaging to a vote count. The Party which is denied the middle has to rely on its base, but that base is not nearly enough to carry a national election unless you also have some way to deny the middle to your opponent. In 2010, the GOP was successful in getting the middle to stay home. There is virtually no chance that the Dems are going to allow them to do that again, especially since the GOP has moved further right than they were in 2010!
And so, as a word of summary on the conventions: we will know how successful they were when we see voter turnout. In a world with perfect voter turnout, the Dems win this election by some insane amount. The middle, and a significant portion of the moderate right is more likely to vote for Obama than for Romney, given the ideological location of the Parties.
It is my opinion that the Dem Convention was far more successful in accomplishing the missions of telling people who there is an election going on and in giving them a reason to turn out. They spent most of the convention talking about issues in an election year which is more oriented toward issues than symbols. The GOP's convention was less successful, simply because it was focused more on how bad a Obama has been, and was very light on actual policy. The dystopic vision of the future, I suspect, will not be enough to prevent people in the middle from coming out to vote. If it does, it will end up keeping people who do not like Obama at home. Unfortunately, those are the same people Romney will need to go and vote for him.
The codes coming from both Parties are clear, but for the GOP, they are not reaching the ears that they are intended for. In effect, they are preaching to the choir, focusing on the people who are never going to vote for Obama anyway. They are not talking to the center, and those are the very voters who the GOP needs to focus on to get them to stay home. It is almost as if the activist wing of the GOP have pulled the Party so far to the right that they can't even talk to the center anymore. Nothing the GOP said during the conventions will decrease voter turnout by a significant amount. That could change over the next two months, but it did not happen at the convention.
Quite the contrary: The GOP may have actually undermined their own turnout at their convention! There is reported frustration and anger at the decision of the Rules Committee to change the rules to prevent Paul's name from being entered into nomination from the floor. It is not known at this point whether that will translate into a significant shift of libertarian minded Republicans away from the GOP ticket, but since a large part of the TEA Party, the right-wing activist insurgency which is responsible for a significant number of GOP misreads listed above since 2009, considers themselves to be libertarians as well, and Sarah Palin has more or less set herself up as a spokesperson for the TEA Party, it is entirely possible that the result could be quite painful in the fall. At any rate, GOP Leaders dismiss this activist anger at their own risk. It must be remembered that a significant number of these people are on the right of the Party's position, and therefore are highly unlikely to ever vote for Obama. But in a turnout game, staying home is a vote for your Party's opponent.
For his part, Obama WAS speaking to the middle. His Party presented a strongly communitarian message, which seems a bit strange in the post-9/11 world, and is certainly not one we have heard for a while. But then again, he has the freedom to do so, given the sharp shift to the right of the GOP. And it is useful in bringing the left, who was wavering on Obama out to support what is essentially a centrist platform with a few liberal Issues on it. In terms of turning out the middle, he asked them to go and vote, and as has been discussed, that request often goes a long way, because voting people equals an Obama victory, given the structural features of this election.
Ultimately, Obama had an easier job: It is far easier to ask for a vote than it is to ask, in code of course, for a person to not vote. Given what we know about elections, and this election in particular, Obama's convention was a success. Nothing he said will get those dissatisfied Paul supporters to come and vote for him. But then again, he doesn't need their vote: all he needs is for them to refuse to give it to his opponent. Anger on the right makes Obama's job that much easier. Again, this could change in the next two months, but as of right now, there will be a decent turn out. And given what we now know about the structure of the election in 2012, this means Obama wins.
-W. Doc Stodden for American Conversation.
