About this Blog

RealClearPolitics HorseRaceBlog

By Jay Cost

HorseRaceBlog Home Page --> February 2008

What's So Bad about the Super Delegates?

Last week I wrote an essay in praise of the Democratic super delegates. I argued that in comparison to the Republicans - they offer real advantages. They serve as a kind of "majority maker" for the party. When a candidate has not won a 3/5ths majority of pledged delegates, the super delegates break the "tie." The Republicans have a "majority maker" solution, too: most states free their pledged delegates after a few ballots. The Democrats' solution is better. Super delegates are free to negotiate whenever they like; they have an interest in finding a candidate who is best for the party; and they have the capacity to engage in the difficult process of negotiation.

However, there are problems with the super delegates. Democrats clearly sense this - perhaps this is driving their desire to wrap the primary up. From a certain perspective, this is a strange preference. Isn't this robust contest helping the party think about its future? Isn't it helping the candidates sharpen their skills? And yet, neutral Democrats would probably be glad to see the race end on Tuesday. From another perspective, this is a highly reasonable thing to desire. While the Democratic process is preferable to the Republican one, it is still inefficient. Democrats have reasons to doubt they can trust the super delegates to do a good job concluding the contest.

The core problem is that the Democrats have empowered the super delegates to break a tie, but they have not empowered anybody to manage the super delegates. There are no rules that demand the super delegates convene and discuss with one another. There is nobody in charge of regulating the debate. There is nothing to punish the super delegates who are small-minded, nothing to reward the big-minded. There are no time restrictions that require them to make up their minds prior to the convention. They are wholly unfettered.

Thus, the super delegates have a great deal in common with a mob. They're a mob of experienced, qualified politicos who care about the party. If the Democratic Party were to be put at the mercy of a mob - this is the mob you'd want. But it is a mob nonetheless. This is why large institutions - like the House and the Senate - have reams of rules governing member behavior. If the members of those institutions are to do their jobs ably, they need a framework for interaction. Otherwise, their talents may be squandered amidst the chaos.

Let's look in depth at one potential problem.

Earlier this week, I argued that each super delegate has a personal interest and a public interest that could factor into their decisions. For instance, each House Democrat has an opinion about who is best for the party. This would be his or her public interest. Each also has a personal interest in being reelected, and this might include placating constituents by voting as they did. This would be his or her personal interest. For many super delegates, there would be no conflict. Their constituents voted the way they prefer. But for some, there will be a conflict - as seems to have been the case with John Lewis. These super delegates face a version of what is known as the dilemma of collective action. Do they pay personal costs for a public benefit, or do they sacrifice the good of the party for their own good?

Let's take a look at a simple, stylized interaction that teases out some implications. Assume there are just two super delegates, both of whom face a conflict between their public and personal goals. Each gets a choice to go one way or the other. Additionally:

- Suppose that if both delegates do what is best for the party - the party appears to be responsible to the public. So, both delegates get a benefit of P. But doing so sacrifices their personal interests, so they pay a cost of -C.

-Both delegates also have an option of doing what is best for themselves. If they do this, they get a benefit of C (regardless of what the other does). However, if one of them chooses to elevate himself above his party, the party will not appear responsbile - and both will pay a cost of -P.

-Accordingly, if both do what is best for the party, both get a benefit of P - C. If both do what is best for themselves, both get a benefit of C - P. If one works for the party and the other for himself, the first gets a benefit of -C - P, and the second gets a benefit of C - P.

These payoffs can be modeled in a two-by-two matrix. One actor "plays" the rows. The other "plays" the columns. Both choose whether to do what is best for the party or best for himself. The actor playing the rows gets the first payoff in each cell, the actor playing the columns gets the second. The generic form of the interaction would look like this:

Game 1.gif

Let's set C = 5 and P = 7. This implies that both candidates enjoy a greater benefit from helping the party than they do from helping themselves. In other words, this is what Democrats would want from the super delegates. What would the interaction look like then?

Game 2.gif

At first blush, one might think that there is a simple solution: both candidates do what is best for the party in the top-left cell. While this is a possible solution (or equilibrium), and it is socially efficient, there is another solution. The bottom-right cell, in which both candidates do what is best for themselves, could also be the outcome of the interaction. This one is socially inefficient. In other words, this interaction could result in the efficient outcome where both support the party or an inefficient outcome where neither does.

This is where some kind of institution could come in handy. That is, some rule or person could alter the payoffs to ensure that the delegates choose what is best for the party. How might this work? Suppose that the super delegates knew that if they support the party, they would be personally reimbursed for their loyalty. This might come in the form of flattering publicity, policy considerations down the line, campaign contributions to offset any electoral danger they may face, or whatever. The point is they know that supporting the party would offer some side benefit that is just theirs. They also know that this benefit is theirs regardless of what the other delegate does. That would change the game to the following by inserting a loyalty (L) factor for delegates who support the party.

That is:

Game 3.gif

Let's set L = 12 and re-run the interaction, retaining C = 5 and P = 7.

Game 4.gif

There is a single solution/equilibrium to this game. Review the options of the row chooser. Note that regardless of what the column chooser decides, he is best off serving the party. If the column chooser goes with the party, the row chooser gets -2 going for himself or 14 going for the party. If the column chooser goes for himself, the row chooser gets -2 going for himself or 0 going for the party. So, his rational move is always to go for the party. The same goes for the column chooser. Thus, both delegates will choose what is best for the party. The difference here is that loyalty factor. What it did was reward the delegates for supporting the party regardless of what the other does. This shifted their strategies.

This sort of personal payoff is actually quite common. If you have ever received a tote bag from PBS, you have received a personal payoff for helping a broader goal. This is what institutions can do. They can offer personal benefits to individuals to guide interactions to the socially efficient outcome. It need not be benefits. We could inverse the above interaction. Instead of a +12 loyalty benefit, we could have a -12 disloyalty penalty. It would have the same effect. This is one reason why the government is empowered to penalize tax cheats.

Mechanisms like this do not exist with the super delegates. They are in an institutional vacuum where there are no rules to govern their behavior, let alone dispense personal benefits if they put the party above themselves. Now, it might be that the preferences of the super delegates are arranged in such a way that an institutional mechanism like a tote bag is not necessary. Indeed, you could reassign the values of C and P in the first game so that both super delegates choose to support the party (e.g. re-run the first game setting C = 0). The preferences of super delegates need not be arranged in a socially inefficient way. But that misses the point. The point is that they need not be arranged in a socially efficient way, either. And if they are not - there is no "tote bag" provision to induce a socially efficient result.

One way or the other, the super delegates will make a choice, and the party is going to have a nominee. The question is how costly it will be to get this nominee. And this is what efficiency is all about - achieving an outcome while minimizing costs.

What are some of the costs that the Democrats will face because there are no rules? So far, we have hinted at the specter of illegitimacy. That is, one candidate wins because he or she is good for particular super delegates, not the party. Another cost could be delay. Because they are not required to do anything until Denver, they might not do anything until then. This would mean that Democrats will face a primary battle that ends six months from now. While I think most would agree that it would do the party no harm to have the primary last another two months, six months would be genuinely harmful. Waiting until Denver also means making a decision under intense, worldwide scrutiny. Party politics is not meant for such a close look. It's inevitably narrow-minded: personal concerns always come to influence what non-partisans think is a strictly public matter. Party deal-making is an illiberal part of any liberal society such as ours, but that does not mean the public will accept it. On the contrary, it will probably turn off the average voter who sees the deal go down on live television.

There are all sorts of other costs. None of them derive from the super delegates themselves. As I argued last week, the super delegate provision is a good "majority maker" solution. The problem is that they are free to do whatever they want.

Personally, I find this lamentable. It seems to me that the Democrats are in the midst of a robust, valuable debate about the future of their party. The fact that it does not involve sharp policy differences is a non sequitur. One need not discuss policy to be substantive. If the RCP average is any metric - it is an argument that neither side has won. And yet, lots of worried Democrats want it over. They doubt the capacity of the party organization to resolve the conflict. They are wise to have these doubts. Because they are unbound by rules of any kind, the fact that the super delegates will break the tie is a disaster waiting to happen.

But why are the super delegates so free? The Democrats have lousy rules that nobody cared to revise in the last quarter century because the best and brightest in American politics don't give a damn about the party organization. This is part of a decades-long trend in American politics. The party institutions have been taken for granted. They no longer play a vital role in daily American political life, so they are left to decay - until we need them. At which point, they are incapable of doing their job.

Americans like to think that strong parties are an impediment to democracy - and so, the weaker they are the better we are. They are wrong. Strong parties are an asset to democracy. The happenings on the Democratic side indicate what can happen when the parties are weak. The Democrats are in the midst of a animated discussion that many of the conversants want to end because the party organization is incompetent. What a shame.

-Jay Cost

Who's the Better Closer?

Last week on Fox News Sunday, Brit Hume made an interesting point:

(Clinton) closes well, you know. Late deciders tend to break for her. So she's got a chance here. The question is will she carry that over into the way she keeps campaigning and will she carry it over to these debates, which are very widely watched.

I have heard this point before. She probably earned that reputation after New Hampshire. Of course, we have had many contests since the New Hampshire primary. Is she still a good closer?

There are two ways we could answer the question. First, we could look at the polls that precede a given contest to see whether Clinton improves at the end, or does better than the polls predict. Unfortunately, this is only practical for states like New Hampshire. For the rest, we do not have enough polling data. Second, we could use the exit polls, which include information on when voters decide. If we exclude each candidate's home states (expecting voters to think about the race differently in those states) - we have 21 states to examine. This is the approach we shall take.

The exit polls ask voters if they decided the day of the election, three days prior, a week prior, a month prior, or sometime earlier. If Clinton closes well, we should expect her to do better among "day of" deciders, as well as perhaps "three days prior" deciders. However, there is a complication. What about a state like Utah? We might expect Obama to beat her among all voters, regardless of when they decided. After all, he won the state by 18 points. So, she might do better with "day of" deciders than with the whole electorate, but we will not pick up on the improvement if we examine the raw data.

Our response will be to "standardize" the results. That is, we won't just examine how Clinton did with a particular type of voters, we'll examine how she did with those voters relative to the state's electorate. For instance, Clinton won 42% of the vote in Alabama. She won 58% of "day of" deciders. So, she did better with them than with the whole electorate - consistent with the "closing well" theory. We'll capture this by dividing her "day of" share by her total share to get a standardized score of: 58 / 42 = 1.38. Conversely, she won 55% of the vote in Oklahoma, but only won 50% of day of deciders. So, her standardized score is: 50 / 55 = 0.91. This procedures enable us to see that she did close well in Alabama, but not as well in Utah.

This gives us a handy decision rule. A value less than 1.0 indicates that the candidate did worse with a group than with the whole electorate. A value greater than 1.0 indicates an improvement. So, if Clinton "closes well," we should expect values of greater than 1.0 for the late deciders. These scores will help us see how each candidate's voting coalition solidifies over time.

We can also develop a measure that compares the candidates to one another. For instance, we know that Clinton lost by 14 points in Alabama. However, she beat Obama among "day of" deciders by 19 points. This means that Clinton did 33 points better with "day of" deciders than with the whole electorate. This is a kind of "Clinton improvement factor." A positive value indicates that the type of deciders in question prefer Clinton to Obama more than the whole electorate, a negative value indicates that they prefer her less.

This gives us three statistics - Clinton's "standardized" score, Obama's "standardized score," and the so-called "Clinton Improvement Factor." We have all three statistics for each type of voter in each state. We average them by type across states and report the results in the following table:

All States.gif

Clearly, Clinton does better with "day of" deciders than Obama. Her score of 1.05 indicates that her share of the vote among them is usually greater than her share of the vote among whole electorate. Obama's 0.88 score indicates that his share is usually less than his share of the whole electorate. We also see that Clinton does about 8 points better relative to Obama with "day of" deciders than she does with the whole electorate. So, we can conclude that: (a) Clinton does better with "day of" deciders than with the whole electorate; (b) Obama does worse; (c) "day of" deciders favor Clinton over Obama more than the whole electorate.

Now, let's examine voters who decide earlier. Obama's position improves markedly. He does better with "three day" deciders, "week" deciders, and "month" deciders than he does with the whole electorate. Clinton does worse with each of these groups. What is more, the negative "Clinton Improvement Factor" values show that these groups favor Obama over Clinton more than the whole electorate. If we move back in time, we see that the "earlier" voters favor Clinton a little more strongly than the "day of" voters do. This is when most of Clinton's voting bloc forms.

All in all, we see evidence to support Hume's conclusion. Clinton does well with voters who decide early. Obama then does very well, but Clinton slowly closes the margin. This general trend holds up when we divide the states into subsets. If we look only at states Clinton won*, states Obama won*, states before Super Tuesday*, states on Super Tuesday*, and states after Super Tuesday* - we see the same pattern: Clinton does well far from the election; Obama does well about a month before to the day of the election; Clinton closes the gap on Election Day.

What could explain this pattern? Many theories could - one that I am partial to takes into account the pre-campaign knowledge voters have of Clinton and Obama. Namely, most voters know a lot about Hillary Clinton. They do not require a campaign to learn about her. So, voters who are "predisposed" to support Clinton make a decision early on that is not overturned once they learn about Obama, which happens during the course of the campaign. On the other hand, voters who are "predisposed" to Obama generally do not know enough about him to make a decision before the campaign begins in earnest. When it does, they learn that he's the candidate for them. What about voters who decide very late? They probably either have paid so little attention to the campaign that it has had no effect on their thinking, or they have been paying attention but remain genuinely ambivalent. Either way, Clinton does better among them because she is so well known; both go to the candidate they know more about. My sense, then, is that Clinton does not so much "close well" - a phrase which implies that the shift is a consequence of her own efforts - as she benefits from having been a national figure for nearly two decades.

As of now, this is just a theory. The exit polls by themselves cannot tell us whether this theory, or some alternative, is more likely. We'll have to wait for the release of more detailed data that probes voter psychology in some depth, like the National Elections Study. I think this particular theory is consistent with what we know about voter information levels as well as Clinton and Obama's relative positions in the electorate. I also think it does a good job explaining the variation we have seen. But this just makes it reasonable. We'll know for sure when we have more data.

A final observation. This argument places less importance on the effect of political campaigns than what many media pundits would accept. Too bad. Campaigns matter, for sure - but the fact is that many pundits fail to understand how they matter because they misunderstand the voters. They seem to believe that voters pay as much attention as pundits do - and therefore shifts in the horse race polls imply shifts by voters who are paying a lot of attention and changing their opinions. While it is true that some voters behave in this manner - what we might call "schizophrenic Jeffersonianism" - the vast majority does not. Note the large number of voters who claim to decide late in the exit polls cited above. This is inconsistent with the idea most pundits have of the voters. It also implies that their analytical emphases - i.e. parsing the day-to-day nuances of the campaign to distill their impact - are misplaced.

This will become relevant as we move forward. We should expect the media to analyze the early part of the general election campaign just as they analyzed the early part of the primary, which is this incorrect way. Expect them to over-estimate the importance of the day-to-day events, only to forget them a few days after they occur, and above all to over-emphasize the value of the opinion polls conducted months before November. When they do that, we should remember these exit polls. Despite the fact that the primary campaign lasted something like 13 months - most voters claimed to wait until the last month or later to make a choice.


[*] States Clinton Wins (9 States)
States Clinton Wins.gif

[*] States Obama Wins (12 States)
States Obama Wins.gif

[*] Pre-Super Tuesday States (4 States)
Pre-Super Tuesday States.gif

[*] Super Tuesday States (13 States)
Super Tuesday States.gif

[*] Post-Super Tuesday States (4 States)
Post-Super Tuesday States.gif

-Jay Cost

Is It Over?

Jonathan Alter has a thought-provoking article in the latest Newsweek. He writes:

If Hillary Clinton wanted a graceful exit, she'd drop out now--before the March 4 Texas and Ohio primaries--and endorse Barack Obama...

Withdrawing would be stupid if Hillary had a reasonable chance to win the nomination, but she doesn't. To win, she would have to do more than reverse the tide in Texas and Ohio, where polls show Obama already even or closing fast. She would have to hold off his surge, then establish her own powerful momentum within three or four days. Without a victory of 20 points or more in both states, the delegate math is forbidding. In Pennsylvania, which votes on April 22, the Clinton campaign did not even file full delegate slates. That's how sure they were of putting Obama away on Super Tuesday.

The key word is "reasonable" - as in Hillary doesn't have a reasonable chance to win the nomination. While I agree that Obama stands a much better chance of winning the nomination than Clinton, I think Alter's conclusion is hasty. If she loses either Texas or Ohio next week - the race will end. Nevertheless, let's assume that she wins both, though not by the large margins Alter says she needs. What happens next?

Neither Clinton nor Obama can expect to win the nomination by virtue of the pledged delegates alone. Obama would have to win more than 75% of the remaining delegates. Clinton needs more than are available. Thus, the nominee will have to fill the gap via the super delegates.

This is critically important. The nominee will be the one who makes a compelling argument to a sufficient number of the 795 super delegates. This is the first reason not to be so quick to declare the race finished. Do we know what these delegates are thinking? We have no survey data on them - nothing that gauges their preferences or beliefs. We can easily track how candidates are doing when their audience is the American public. That's what opinion polls are for. We have nothing of the sort for the super delegates.

Even though we do not know what these super delegates will decide, we can draw a general outline of how they might decide. Doing so should demonstrate just how complicated matters could become. Their decision will hinge upon their answers to two general questions:

(A) Which candidate is better for the party?

(B) Which candidate is better for me?

Super delegates are free to answer these questions however they like. They are also free to weigh the answers however they like. They can be selfish or sefless - whichever they prefer. That's what makes them super. This adds another dimension of uncertainty. Not only do we have few indications about their thought processes, we know that their thought processes are unconstrained by any party rules.

Let's take a closer look at each question, beginning with the "What's in it for me?" query. In the 19th century, presidential candidates could use patronage to offer all kinds of perquisites. They cannot do this any longer. So, personal interests will hinge on answers to the following questions:

(1) Which candidate do I prefer?

(2) Which candidate do my constituents - whose support I need for reelection - prefer?

Question (2) will matter less to the super delegates who do not hold elective office. However, it could matter a great deal to those who do, and answers to it could be quite different.

Suppose that House and Senate Democrats, all of whom double as super delegates, decide that they shall reflect the will of their constituents. How would they vote? It depends. Obviously, senators would follow their states. But what would House members do? They might ask:

(a) Do I follow my state's vote?

(b) Do I follow my district's vote?

Their individual answers could have a sizeable cumulative effect. If all Senate and House members follow their state results, Clinton would have 108 super delegates (so far) to Obama's 88. If, on the other hand, House members follow their congressional districts, Clinton would have 86 to Obama's 110. That's a 44 delegate swing - over what is a tiny distinction.

Obviously, some will follow their districts, some will follow their states, and some will support the candidate they personally prefer. Some might not even be moved by personal concerns - and will instead support who is best for the party.

But who the heck is that?

It seems to me that there are at least two ways for a delegate to consider who is best for the party. They could ask:

(1) Which candidate is more able to defeat John McCain?

(2) Which candidate is the legitimate choice of Democrats nationwide?

These two questions are probably not exclusive. The candidate who is perceived to be the choice of the Democratic electorate will probably be better positioned for the general election. That being said, we can still look at each question individually. It is here that the candidates will begin to make real arguments. Each delegate is going to have to make and weigh decisions about their personal interests by themselves. Where the Clinton and Obama "spin" machines can have their greatest effect on the delegates is persuading them that their particular candidate is best for the party.

What could each candidate argue vis-à-vis McCain? Obama can point to his lead in the head-to-head polls as well as the "Obama-mania" that has overtaken part of the country. He can assert that his supporters are more dedicated, and will give him a better donor and volunteer base. Clinton has a good argument, too. She can reference the old adage that a bird in the hand is worth two in the bush. Sure, Obama enjoys this enthusiasm now, but it only matters if it is there in November. Will the Republicans tear him down the way they "Swift-Boated" John Kerry? Clinton can argue that they won't be able to do this to her. They have been trying to no avail for sixteen years.

The second question - who is the "legitimate" candidate? - will be the most important. As of now, Obama has a clear advantage. He leads in the delegate count as well as all tabulations of the popular vote. This is why Texas, Ohio, and Pennsylvania are must-wins for Clinton. She has to make a compelling argument that she is the legitimate choice of the party. This will require victories in all three.

Assuming that she wins them, what would an argument for Clinton as the legitimate nominee look like? It probably will not include the claim that she leads the pledged delegate count. Alter is correct: she probably won't. Her objectives must be to close the gap by a good amount (victories in the big states should help), and diminish the importance of the count. There are several ways she could accomplish the latter task. First, she could say that Obama's strength is in red states that John McCain will carry in November, while her strength is in the heart of Democratic territory. Second, she could say that she wins Democrats while Obama wins Independents - and the delegate allocation process does a poor job excluding the latter from a process that should belong to the former. Third, she could argue that the delegates for Michigan and Florida should be seated - that the nominee of the party should not hinge on a Carl Levin power play that backfired.

Fourth, and most important of all, she could attack the fairness of the caucus system. This will be her best bet, I think. Obama has won most of the caucus states overwhelmingly. Clinton could assert that the caucus favors Obama by unfairly excluding voters who happen to favor her - namely, "downscale" Democrats who cannot take off work to attend and elderly voters who are unable to. Clinton will have some evidence to buttress this claim. The Washington state caucus allocated 68% of the state's delegates to Obama on February 9th. Ten days later, on February 19th, the state held a non-binding primary in which Obama won 51% of the vote. Texas might yield a similar result. If Obama beats Clinton in the caucus, and she beats him in the primary - Clinton can argue that the caucus system unfairly skews toward him.

None of this will matter, however, if Clinton does not have a lead in a nationwide vote count. I cannot see her arguing for legitimacy without this. RCP is keeping track of the vote leader by three metrics: including Florida and Michigan, including Florida, and excluding Florida and Michigan. Clinton needs to be the leader by one of these. Her legitimacy case will be stronger if she leads by more than one. In that case, she could try to paint Obama as a Democratic version of George W. Bush - somebody who lost the popular vote but nevertheless "won" by virtue of the quirks of an outdated, unfair system that is still around because nobody cared enough to get rid of it before it created trouble.

Can she take a popular vote lead? Possibly. The average turnout in the primary contests that have already occurred has been 12.4% of the total population of those states. If we assume that the remaining primaries will have the same turnout - we can make the following observations:

(a) She would have to win more than 54.7% of the remaining vote to take a lead in the count that excludes Florida and Michigan.

(b) She would have to win more than 53.2% of the remaining vote to take a lead in the count that excludes Michigan.

(c) She would have to win more than 51.5% of the remaining vote to take a lead in the count that excludes nothing.

This is why Texas, Ohio, and Pennsylvania are important. Combined, they make up 60% of the remaining population left to vote. If Clinton wins all three by solid margins, her burden will be greatly diminished.

Of course, Obama is currently much better positioned for a legitimacy argument. He has all three vote leads - and he is closing fast in Texas and Ohio. Clinton can only threaten him if she finishes strong enough to eliminate one of the popular vote gaps. Even then it might not be enough. It will be hard for her to win 54.7% of the remaining vote. If she does not, Obama will have at least one lead. If the best she can do is win 51.5% of the vote, Obama will be well-positioned to argue that including Michigan is completely unfair because he was not even on the ballot. What is more - the three metrics necessarily exclude a few caucus states like Iowa that do not report actual votes. This would help Obama rebut Clinton's claim to be the "choice of the voters" if she were to take the lead.

Obama can respond to most of Clinton's other arguments. For instance, he could argue that the pledged delegate count is what matters, and that the super delegates should not presume to alter the outcome of the primaries/caucuses. He could assert that Clinton's anti-caucus argument is just the bellyaching of a candidate who failed to prepare adequately for the contests on and after Super Tuesday. This could even pivot into an electability argument. Sure, Clinton cannot be "Swift-Boated," but she was unprepared for the primary process. Shouldn't this worry the super delegates?

What should be clear from all of this is that, if Clinton does well on March 4th and in Pennsylvania next month, things could get very complicated. Let's review our outline of the questions the super delegates will answer. Remember that each must formulate his or her own answer, and must weigh the answer to each question for himself or herself.

(A) Which candidate is better for the party?

(1) Which candidate is more able to defeat John McCain in November?
(a) What are the chances Obama's reputation will be diminished by Republican attacks?

(b) What are the chances Clinton will under-prepare for the general election?

(2) Which candidate is legitimate choice of Democrats nationwide?

(a) Who has a lead among the pledged delegates?

(b) Does it matter what types of states and voters each candidate has won?

(c) Is the caucus system sufficiently democratic?

(d) Should I care about the popular vote leader...

(i) ...including Michigan and Florida?

(ii) ...including Florida?

(iii) ...excluding Florida and Michigan?

(B) Which candidate is better for me?

(1) Which candidate do I prefer?

(2) Which candidate do my constituents - whose support I need for reelection - prefer?

(a) Do I follow my state's vote?

(b) Do I follow my district's vote?

This set of questions is surely not comprehensive. There are others that will factor into each delegate's decision. This will make their decisions all the more complex.

You probably have opinions on all of these points. So do I. But here's the kicker - our opinions don't matter. We're just spectators. It's up to the super delegates, and we have no idea how they will decide.

I think it is hasty to say that Clinton lacks a "reasonable" chance to win the nomination. If she wins Texas, Ohio, and Pennsylvania - the race will hinge upon how each super delegate answers and values these questions. I'm not saying that I favor Clinton to win. At this point, I don't. She stands a chance if she wins the three big states that remain. He stands a chance regardless. This tips the scales in his favor. He also has an advantage due to the pledged delegates - the more of those you have, the fewer super delegates you need, the less pressure there is for you to argue a case. Nevertheless, Clinton still has a reasonable shot if she can win next week.

-Jay Cost

Momentum in the Democratic Race

Is there momentum in the Democratic race? If there is, when did it begin?

This is a question that has been on my mind this week. On Wednesday, I argued that there is evidence that Obama developed momentum in Wisconsin. I also asserted that this did not seem to have been the case in Virginia and Maryland. It was this latter claim that went against the grain. Many thought that Obama did indeed develop momentum for the Potomac primary. I did not.

How could we determine whether momentum has indeed taken root? The most direct way would be to use the exit polls to look at key demographic groups. If Clinton seems to be doing worse with groups she was once strong with - we might conclude that momentum is playing a factor. Many commentators concluded this after Maryland and Virginia. They found that Obama was doing better among Catholics, union workers, women, and "downscale" voters. Accordingly, they declared that he was benefiting from momentum.

But hold on. Demographic groups are not hermetically sealed off from one another. There is a lot of overlap. Any discussion of voting blocs would have to take this into account, especially in a state where one candidate's strongest group is in large supply. In such a situation, the other demographic indicators might be "overwhelmed." This seems to have been the case in both Maryland and Virginia. In fact, Obama had two of his best groups in large supply: African Americans and wealthy voters. This makes a big difference. We need to remember that they too can be union workers, women, etc. Their strength at the ballot box might make it look like Clinton is suddenly losing her groups, but these were never actually her groups.

Compounding this is the fact that there probably was not a great deal of overlap between the two. Obviously, they are not mutually exclusive. Many African Americans make $100,000 or more. However, whites in Maryland make 39% more than African Americans in Maryland; in Virginia, they make 57% more. So, while there is surely some overlap, it stands to reason that African Americans plus wealthy whites made up about half the vote in both states. So, it seems likely that momentum was not a major factor. A careful reading of the exit polls indicates that Obama probably did not "steal" votes from Clinton, but won on the strength of his typical voters.

Of course, we cannot know for sure. As it is, the exit polling data is just not detailed enough to answer our question directly. What to do? One solution is to use more detailed cross-tabulations of exit poll results. Unfortunately, the media does not provide them. We could also wait. We should know in a year or so; when the PhD's who work for outfits with money to do detailed polling publish their write-ups on the Democratic race, we'll get some answers.

A practical, if imperfect, solution for right now is to combine this limited data set with another limited data set. Namely, we can use the exit polls in conjunction with countywide vote returns. This cannot give us a definitive answer (ultimately, momentum has an effect on individual voters - so we would ultimately need survey data to know for sure). However, it can give us something better than what we have now.

Unfortunately, we cannot use the returns by themselves. The biggest problem is with something known as the ecological fallacy. Specifically, our knowledge of how a county voted cannot give us direct knowledge of how individuals in that county voted. This is a real problem. We would have to make an assumption that, for instance, the observation of a poor county voting for Clinton implies that the poor individuals in that county voted for her, too.

This is where the exit polls can be helpful. If the exit polls imply that "downscale" voters are supporting Clinton, this assumption becomes much more reasonable. Conversely, we can use the aggregate voting results, imperfect as they are, to "firm up" the tentative read we get from the exit polls. The idea here is that, while both types of data present inferential problems, we are on fairly solid ground if the exit polls and the aggregate results point in the same direction.

Before we dive into the countywide returns, however, we have to specify matters a bit more. In particular, we face a problem with racially diverse counties. Their returns cannot help us. Consider, for instance, King and Queen County, Virginia. The median income of whites there is about $41,000. Presumably, this favors Clinton. On the other hand, African Americans constitute 35% of the population. This favors Obama, who the county by 44 points. Was his win due to momentum? Or was it due to his voting bloc overwhelming Clinton's bloc? We can never know. So, this type of county is of little use.*

So, we will look at counties where the African American population is small, say less than 5%, to see whether they varied their support according to median white income. Our expectation, following the exit polls, is that because momentum did not seem to exist - Clinton will do well in the "downscale" counties, and Obama will do well in the "upscale" counties.

Let's look at Maryland first. There are three Maryland counties with an African American population of less than 5%. Clinton won Cecil and Garrett counties by more than 20 points. She won Carroll County by just 3 points; however, this is explicable by the fact that the median white income of Carroll County is more than $60,000 per year.

In Virginia, Clinton generally did well in the homogeneously white counties. We find a tight relationship between income and countywide returns, and it is in the direction we expect. The following graph tells the tale. It examines Clinton's margins of victory in homogeneously white Virginia counties according to median white income.


First, note the distribution of the data points. It indicates a strong relationship between income and Clinton's margin of victory.* Note also the small number of homogeneously white counties. Virginia has more than 70 counties (or cities) - and about 40 of them have African American populations of some size. Clearly, Clinton had an uphill battle in this state.

Our real interest is in the four quadrants divided by the black lines. The horizontal line separates Clinton wins (top) from Clinton losses (bottom). The vertical line separates counties where the median income is less than $40,000 (left) from counties where the median income is more than $40,000 (right).*

The array of the data points on the graph is roughly consistent with what we might expect.* We see a large number of observations in the top-left quadrant, and most of the observations on the right-hand side are in the lower-right quadrant. In other words, Clinton is generally winning the counties where voters make less than $40,000; Obama is generally winning the counties where whites make more than $40,000. Both candidates "take" just a few counties that "belong" to the other. Again, by itself this tells us nothing because it provides no indication of how individuals are behaving. However, taking it in conjunction with the exit polls, we have a good indication that momentum was not driving the results in the Potomac primary. Instead both candidates were doing well among their core groups; Obama's groups happened to have been more populous.

Now, let's turn our attention to Wisconsin. As we discussed on Wednesday, the exit polls indicate that Obama may indeed have developed momentum. Specifically, they showed that Clinton seemed to do worse-than-expected among lower income whites. But, again, we cannot be sure because the exit polling cross-tabulations are limited. The aggregate data might be able to supplement it. The following chart reviews returns by median white income for homogeneously white counties.


First, notice that income and county returns still have a reasonably tight relationship, one that moves in the direction we expect (i.e. Clinton does worse as income increases). Second, it appears that Obama benefited from the fact that homogeneously white counties in Wisconsin tend to be wealthier than those in Virginia - we have many more observations to the right of the horizontal line than we had in Virginia. Not surprisingly, they are in the lower-right quadrant. Third, when we look at the four quadrants, we see a pretty significant shift. Obama wins a much larger number of poor counties than he did in Virginia, where Clinton took most of them. We can see this in the large number of observations in the lower-left quadrant. These are counties that, if Wisconsin were following the pattern in Virginia, would be in the upper-right and going for Clinton. Taken in light of the exit polling data, this is strong evidence that there was a shift among "downscale" whites in Wisconsin.

I think this shift is hard to account for without hypothesizing a momentum effect. Consider the following. We know that the Hispanic population, Catholic population, and union population are important factors in how each candidate performs. The difference in the Hispanic population between Virginia and Wisconsin is negligible (4.7% in Virginia to 3.6% in Wisconsin). There are big differences in the Catholic and union populations between the contests, but these should have favored Clinton. The explanatory variable that is not accounted for above that favors Obama is racial homogeneity. It is possible that lower income Wisconsin voters are more amenable to Obama because race is perceived differently in Wisconsin than it is in Virginia. However, I do not think this could explain the whole of this shift. Momentum might have been the critical ingredient.

All in all, the exit polls and the aggregate vote results seem to point in the same direction - a shift in Wisconsin, but not in Virginia and Maryland. Momentum may have been a factor. Again, the best way to approach the question is as we have here - through a combination of the exit polls and the aggregate vote data. In both states, they move together. Neither gives a definitive indication, but both give the same indication. It might be that the indications of the exit polls and the aggregate returns are both wrong, but this is unlikely.


[*] This is not to say that voters in these counties are not following the patterns we have hypothesized. They may be. The trouble is that we just cannot make any determinations because voters in counties like these are probably going strongly in opposite directions.

[*] However, the relationship is tighter or weaker depending upon Clinton's margin of victory. The clustering is very tight in the top left, but the data points slowly disperse as we move to the bottom right. This property is known as heteroscedasticity, and it would complicate drawing a naïve prediction of a county result based on income. Our purpose here is more modest; so, we need not concern ourselves with this.

[*] Why $40,000? The Census Bureau only releases income data for the decennial census. Accordingly, the county-level income data is eight years old. To account for this, I have adjusted the line on the graph for inflation. Specifically, this year's exit polls tend to show Clinton winning white voters who make less than $50,000. In 1999 dollars, this works out to about $40,000.

[*] The key word in this sentence is "roughly." There are three factors that complicate any simple inferences.

(1) As indicated above, we are dealing with inflation-adjusted income statistics. The difficulty here is that we must assume that voters in these counties are making the same real income now as they were eight years ago. This might not be the case.

(2) Party identification and income might be related. For instance, Democrats make less than Republicans in these counties. In this case, the median income for Democrats in a given county would be lower.

(3) Factoring out African Americans does not eliminate our inferential problems. It only reduces it. Specifically, simply because we know the median income of a county does not mean that we know how individuals are voting. It might not be the case that a poor county has gone to Clinton because the poor voters voted for her. We discussed this problem above, and decided that our solution would be to look at the aggregate returns in conjunction with the exit polls.

-Jay Cost

How Obama Won Wisconsin

Hillary Clinton suffered a stinging blow last night, losing Wisconsin by 17 points. What is most worrisome for her is that Obama seems to have broken into several of her core voting groups. This is the first real evidence of momentum we have seen on the Democratic side.

After the Potomac Primary last week, some argued that Obama had already begun to build momentum because of his large victories in Maryland, Virginia, and the District of Columbia. I thought this talk was hasty. Given the large number of African American voters in each contest, and given that white voters in all three primaries were quite wealthy - Obama's sizeable victories did not come as a surprise. In particular, 37% of Maryland voters and 30% of Virginia voters were African American; 39% of Maryland voters and 39% of Virginia voters claimed to make $100,000 or more per year. So, it is hard to argue that Obama's success was due to him peeling off portions of the Clinton coalition. What seems more likely is that he won handily because his best voting blocs were in good supply that day.

The same cannot be said for Wisconsin. Just 20% of Wisconsin Democratic voters claimed to make $100,000 or more per year. Only 8% of the electorate was African American. So, Wisconsin did not favor his strongest groups so heavily. Nevertheless, he still had an outstanding day. Consider the following chart, which uses the exit polls to compare Obama's margin of victory with key groups in the non-southern states to his performance with those same groups in Wisconsin last night.

Obama Margin of Victory.jpg

So, for instance, Obama won white males in the non-South by 8 points. Last night, he won them by 26 points, yielding a net increase of 18 points.

As you can see, Obama enjoyed significant expansions in three of his four strongest demographic groups. Though it appears he did worse among white Protestants - the difference between the two is possibly due to statistical sampling error. So, all in all, Obama did not do significantly worse with any of his groups - and with most of them he did significantly better.

Meanwhile, he was able to peel away portions of Clinton's core electorate. To appreciate this, consider the following chart, which compares Clinton's margin of victory over Obama in the non-South among the groups that have typically favored her.

Clinton Margin of Victory.jpg

These numbers tell the tale succinctly. Clinton suffered significant losses across many of her core constituencies. White women, Democrats, union workers, downscale voters, and white Catholics all drifted to Obama last night - some so much that Obama actually won them.

Let's cross-reference the exit polls with a look at the actual vote returns. The following chart reviews Wisconsin counties according to median white income. Specifically, it divides 71 of Wisconsin's 72 counties into three tiers according to wealthiest, poorest, and middle. It then looks at Obama's performance in each tier - calculating his average share of the vote across the counties in the tier, and then his share of the vote for the whole tier. [Note that no returns were available for Clark County by the time I had to put this chart together. Most counties had complete returns, and the ones with partial returns were all mostly reported. So, these numbers will probably be different when all the returns are in, but only slightly so.]

Obama Performance by Income.gif

As you can see, Obama did better as whites get wealthier. However, he does very well across all tiers. This confirms what the exit polls indicate: Obama retained his own strengths while cutting into Clinton's strengths. We would have expected her to win the poorer counties, but in fact she lost them.

I was also interested in looking at vote returns according to population density. If Clinton was doing well with economically "down scale" voters - it stands to reason that her strength in a state like Wisconsin would be with counties that are sparsely population. Obama, on the other hand, would do well with counties with dense population centers. This dichotomy probably would not work in a state with large minority populations because of Clinton's strength with Hispanics. However, in a state like Wisconsin, we might expect Clinton to do better in rural areas and Obama to do better in urban ones.

I tested this hypothesis by again dividing counties into three tiers, this time according to population density. The following chart reviews the results.

Obama Performance by Density.gif

Again, we see a similar phenomenon. While Obama did best in the dense counties like Dane and Milwaukee, he also won a majority in the middle-density counties and the sparse counties. There seems to have been no urban-rural divide. The state went to Obama regardless of population density. This is another indication that Obama cut into strengths that Clinton has enjoyed in previous states.

Is this a momentum effect? The word "momentum" has been tossed around way too much this cycle, which is funny because prior to tonight there has really been little evidence of momentum at all! So, what of last night? Are these bona fide expansions in his voting coalition, or was Wisconsin following the same pattern that the previous states have followed? It is impossible to be sure. On the one hand, I have found that Obama does well with whites in states where there are few African Americans. Duke's Brendan Nyhan has found the same trend. White voters in homogeneously white states seem to be more amenable to him than whites in diverse states. Wisconsin is a state with few African Americans. This probably gave Obama a boost last night. If this was a major factor - perhaps some of this apparent momentum effect would "wash out." On the other hand, could the racial homogeneity of Wisconsin alone really account for these huge shifts? That's a tough pill to swallow.

Unfortunately, we simply do not yet have enough data to give a definitive answer. Ideally, we could answer the question if we had enough observations to make a prediction of how Obama should have done in Wisconsin, given the results in past states. We could then compare the actual results to the prediction and see whether the difference between the two is statistically significant. That difference, if it is significant, could be a momentum effect.

My sense is that a momentum effect of indeterminable magnitude emerged last night. That is, Obama's victories - most recently last week in the Potomac Primary - contributed to the size of last night's victory, though he would have done very well if there had been no momentum. This is the first Democratic contest where I think a case like this can be made. Until last week the typical ebb-and-flow of each candidate's demographic strengths could probably account for the actual results.

This surprises me a little bit. As I noted last week - it is rare for a front running candidate like Obama to develop momentum like this. It is not unheard of (e.g. George H.W. Bush had something like this in 1988), but it is rare. Usually, candidates like Obama are "launched" with a splashy win in New Hampshire. That is different from what seems to be happening here. Obama seems to have built upon his latest wins.

If there is a sizeable momentum effect, Clinton should be very nervous. Demographically, Texas has a lot in common with California, except that there are more African Americans. This bodes well for Obama - and if momentum is now in the equation, Clinton could be in real trouble. If you take the margin of her California victory, factor in the larger African American base, and factor in a 5 to 7 point shift in the white vote to Obama - that win might become a loss.

-Jay Cost

What's So Great About the Super Delegates?

With the Democratic race as close as it is, analysts are paying attention to the so-called "super delegates" - namely, the 800 or so party leaders who get to vote however they want, regardless of any primary or caucus result.

What should we make of these delegates? Many analysts seem to approach them with a critical, negative assumption. I think that is presumptuous. I'd like to approach them from a neutral starting point to delineate the strengths and weaknesses that they bring to the nomination process. Tomorrow, we'll look at the weaknesses. Today, let's review the strengths.

At the outset, we should note the super delegates could only be a factor when no candidate wins an outright majority with pledged delegates. This indicates one way the nomination battle differs from most American elections: a plurality of votes is insufficient for victory. A corollary is the Electoral College. If no candidate wins an outright majority of electors - the House of Representatives decides the race.

Just as the Constitution uses the House in the absence of a majority winner, the Democrats use the super delegates. This demonstrates the need for a contingency plan in a majority rule election. Most of us usually never think about this because, outside Louisiana, American elections are decided by plurality rule, i.e. where the winner is simply the person who gets the most votes. There are costs and benefits to both rules. Obviously, with majority systems you need some sort of mechanism to sort out the mess when nobody wins a majority. Plurality systems do not require this - and so they may seem less arbitrary. On the other hand, plurality systems can and often do yield a "perverse" result - a candidate whom most voters opposed nevertheless wins.

One might respond that the Republicans have no such contingency plan for their nomination - even though their nominee is also selected by majority rule. So, why must the Democrats? In fact, the Republicans do have a plan. Theirs is just informal. The Republican solution is that most delegates become like super delegates after a few rounds of balloting. Some Republican delegates are obligated to their candidate as long as he is in the race - but most of them are free to vote their consciences after a few rounds.

This offers a different way to understand the super delegates. Perhaps they seem more reasonable than they first appeared. Every majority system must have some kind of contingency plan for when nobody has won a majority. If we accept the legitimacy of the majority requirement (and why wouldn't we?), we necessarily accept the need for a "majority maker" clause. The Constitution uses the House. Louisiana uses a run-off. The Republicans create de facto super delegates out of the rank-and-file. The Democrats give that power to party leaders.

So, the real question is how good is the Democratic solution? I think it has several advantages over the Republican one.

First, I think that if you are going to make any type of delegate "super" - it is best to make it the party leaders. They are most likely to have the interests of the party as a whole close to heart. To appreciate this, imagine what would happen if there were a knockdown, drag-out fight between McCain and Romney. The only concern on the minds of McCain delegates would be getting the nomination to McCain. Ditto the Romney delegates. But who is looking out for the party? Which delegates will calmly recognize that the elevation of their man would require a nasty battle that might do damage to the party's prospects? Neither. The McCain delegates would probably prefer a nasty floor fight that McCain wins to a cordial process that he loses because their paramount concern is the success of their candidate. Ditto the Romney delegates.

Of course, it is possible that no Republican delegates would behave in such a "narrow" fashion. The trouble for the GOP is that it is also possible that all of them would. This chance is much reduced on the Democratic side because party regulars are intimately involved. They are more likely to care deeply about the party's broader interests. Thus, they can help broker a deal that brings peace to the convention, which is good for the party.

Second, party regulars are more "qualified" to handle a situation in which a deal must be brokered, and making them super delegates gives them the power to do it. To appreciate this, consider the relationship between the House and the Senate. The House was originally envisioned to be the body with the direct link to the public. Accordingly, all tax bills must originate in the House. On the other hand, the Framers gave the Senate functions like ratifying treaties, and confirming officers of the executive and judicial branches. The Senate was ideally to be populated with wise men who could negotiate situations that might be too sensitive for the more raucously democratic House.

We can see a similar logic differentiating the pledged and super delegates. Pledged delegates are certainly politically active - but they need not be professional pols, schooled in the ways of political horse-trading. The super delegates, who are professional or retired political operatives, are well-suited for the negotiating what would have to happen if no candidate wins a majority of delegates. This kind of wheeling and dealing might not be as easy as it appears. If no candidate has a majority, somebody is going to have to switch their votes. They are not going to do that out of the goodness of their hearts. They will need some kind of consideration in return. This type of situation requires a deft touch, which a professional politician is more likely to possess. If the job was left entirely to rank-and-file delegates, all of whom are passionately committed to their candidate - it is easy to envision one faction alienating or offending another due to their inexperience at negotiation or their enthusiasm for their cause.

Third, the super delegates are free to coordinate well in advance of the convention - whereas pledged delegates are not. To appreciate the value of this, it is again instructive to compare the Democrats to the Republicans. Modern communications technology has altered the role of the convention. Party members no longer need to gather in one place to determine whom they prefer. This, in turn, enables them to coordinate long before the convention. That being said, most Republican delegates face an impediment to early coordination that the Democratic super delegates do not. Most Republican delegates are usually bound to one candidate or another for at least a few ballots. This surely complicates pre-convention deal making. It would be hard to work out an arrangement that cannot take effect until the third (maybe even the fourth) ballot. Participants in the deal might have to vote insincerely for the first few rounds to prevent an undesirable result. This enhances the likelihood that a mistake, misunderstanding, or just a shifting situation will kill the deal. It is just a level of complexity that gets in the way of a quick, easy resolution. The Democratic super delegates do not face this problem. They could begin working on a deal right now that could be put into effect on the first ballot.

These three traits are all strengths because they enhance efficiency. That is, they help secure the party a nominee while minimizing the "transaction costs" inherent to any kind of mass gathering. This is an important characteristic for a process like this. Being the nominee brings nobody - be it the candidate or the rank-and-file - any real benefit. It is simply a means to an end, which is victory in November. The more efficient the nomination process, the better positioned everybody in the party is for the general election. Nobody in the party has an interest in a tortuous convention that takes a long, painful time to find a nominee. The super delegates serve as a preventative measure. By virtue of their interest in the health of the party, their experience with deal making, and their freedom to maneuver - they can help the party avoid a messy floor flight.

All in all, I think the Democrat's "majority maker" solution is much more efficient than the Republican's. Of course, it might not seem to matter in a year like this. This is not a three-way race; thus, one candidate might be able to win a bare majority of the pledged delegates. However, it all depends on how strict the definition of "majority" is. I think the Democrats have a stricter definition of "majority" than the Republicans, and this is a prudent move.

To win the nomination without the intervention of the super delegates, a candidate must win 62.25% of all pledged delegates. [i.e. He or she must win a bare majority, or 2,025, of the total delegates. There are 3,253 pledged delegates; thus, a majority that comes through only the pledged delegates would be 2,025 / 3,253 = 62.25% of pledged delegates.] So, winning the nomination through the primary/caucus route is more like getting a bill passed in the Senate than in the House. You need a super majority of pledged delegates.

The prudence of this requirement can be seen if we imagine the process without the super delegates: all a candidate must do is win a bare majority of the pledged delegates. Now, factor in all of the quirky twists and turns we have seen this cycle. Start saying aloud all of those questions you have been whispering for a few weeks. Should we seat the Michigan delegates? What about the Florida delegates? Is the caucus system appropriate for selecting delegates, or should we stick with the primary system? What happens if a candidate wins the big states but loses the little states? What happens if a candidate comes on strong at the end, but does not win enough delegates? What if the party as a whole starts to feel buyer's remorse after a candidate has won a bare majority of pledged delegates? There are all sorts of ways in which a simple majority of delegates might not be sufficient to give the impression of legitimacy, which is a very important value for a nominee to possess. By requiring a candidate to win 62.25% of the pledged delegates - you greatly reduce the likelihood that all of those lingering, partisan-twinged (don't forget the Republicans are watching!) questions could influence perceptions of legitimacy.

And what happens when there is less than this super majority? The decision is left to the super delegates, who - by virtue of the three aforementioned qualities of interest, capacity, and freedom - are in the best position to disentangle which candidate is the "legitimate" nominee. The super delegates are majority makers if we take the definition of majority a bit more strictly. This, I think, points to a counter-intuitive advantage that they offer. Whereas most analysts seem to assume that they are inherently de-legitimizing - I think the opposite is true.

Nevertheless, there are problems with the super delegates. While they might be an improvement over the Republican way of doing business, they present some difficulties that might cause real headaches for the Democrats. We shall discuss this tomorrow.

-Jay Cost

Some Reflections on Polling in the Primaries

The polling has been bad this primary cycle. Last year's national polls were wrong - for both parties. The late polls in New Hampshire were wrong. Even in a state like Virginia, where McCain was supposed to win by a huge margin, he only won by a modest margin. What is more, pollsters have disagreed in state after state. One predicts California going to Romney. Another predicts it going to McCain.

What is going on?

There are surely many answers to the question. I'd like to suggest one I have not seen discussed in depth.

Let's start by taking note of an observation that political scientists have made since the 1950s. That is, average voters do not pay much attention to politics. This is a hard pill for political junkies to swallow, but swallow it we must. Indeed, I think most of the many inferential errors of inside-the-Beltway pundits can be chalked up to their false assumption that voters pay as much attention as they do. They need to get over this. It is just false.

We move from this starting point to a question: how do voters pick their candidate despite these low levels of information? Most researchers will tell you that they make use of cognitive heuristics - mental shortcuts that help them make decisions amidst uncertainty. Uncertainty is the consequence of inattention. Voters simply do not know that much about what is going on - but they nevertheless make a vote choice. Their mental shortcuts are tried-and-true ways of making good decisions despite this uncertainty.

Think of it this way. Political junkies might know the in's-and-out's of each candidate's health care proposals. They can thus make decisions about which they like best. However, the average voter does not have this kind of information. Yet, when he gets to the voting booth, he gets the same choice that the junkie does. The shortcut is his way through the uncertainty.

This leads us to another question - what serves as the shortcut? The answer for virtually the entire country is partisan identification. Upwards of 90% of the nation has some kind of party affiliation. This is despite the polls that identify the size of the independent vote as larger. It is not. There is, for sure, some portion of the country that tells the pollsters they are independent - but when we look carefully at them, we see that most of them lean to one party or the other.

Not only is party identification held by almost all of us - it is an incredibly precise predictor of vote choice. Republicans will almost always go 90/10 for the Republican candidate. Democrats will do the same for their candidate. Most of us have a party identification, and most of us rely on it quite heavily.

What does that mean in a general election? If partisanship is a near universal feature that is incredibly powerful, the preferences of most voters are anchored throughout the campaign - even though they are paying very little attention. They do not have to pay attention to know whom they will vote for. Accordingly, we will see the polls vary only a little bit throughout the campaign. Oftentimes, they will break in late October or even early November. However, the magnitude of the break will be relatively modest. This is not to say it will be inconsequential; just a few point swings in either direction could make a difference in many states. They might swing by +/- 10 points during the whole cycle - but this is paltry compared to some of the massive swings in this primary cycle.

A big reason for this stability is partisanship. As I said, it serves as an anchor. That is a good metaphor for it. Partisanship anchors preferences, keeping them from swaying, drifting, or listing wildly during the campaign.

In a primary campaign, voters must choose among candidates who are all of the same party. Partisanship therefore does not enter into their decisions. It is a non-factor. I think this might be inducing the wild swings in the polls. The polls are varying because the voters are; the voters are varying because their partisanship is not stabilizing their preferences.

Of course, primary voters tend to pay more attention to politics than general election voters. This probably makes them more able to make decisions without the use of their partisanship. Nevertheless, they still pay a price for not being able to use it. To say that primary voters are better informed than general election voters is not to say that they are well informed, or that they behave how the media implicitly assumes they do (i.e. carefully following every speech, parsing every sentence, keeping a constantly updated evaluation of the state of the race, etc). They do not do this.

It thus should be unsurprising that candidate personalities are so influential in voters' decision-making processes. How else do you make determinations when party distinctions are non-existent? Candidates often try to create clear contrasts, but these usually amount to making mountains out of molehills. The average voter is not really paying much attention, anyway. Thus, they have to go by their personal evaluations of the candidates.

This is not to say that vote choices are random. Clearly, there has been a regular pattern to the early contests. Certain types of voters obviously prefer certain types of candidates. The point is that without partisanship, personality is what makes the difference. And voters do not start to take careful note of personalities until late in the cycle. Consider that 48% of New Hampshire Democrats claimed to make their choice in the last week of the campaign. From a certain standpoint, that is incredible. If you think about all of the attention political junkies have paid to this race since last January - it is almost unbelievable to think that voters would not have decided months ago. But, if we put ourselves in the shoes of the average voters, and try to recreate their thought processes - it makes a lot of sense. Their partisanship cannot serve as a quick, easy guide. Thus, they have to take a good, long look at the candidates as people. Given their typical inattention to politics, the time when this happens is the last week or so.

This might explain the wide variability of the primary polling. Because they have not been anchored by partisanship - voter opinions have been unstable for most of the cycle, up until the very end when we are wont to see a massive break in one direction or another. The "error" in the polls might simply be a reflection of public indecision. For that matter, Clinton's massive lead through most of last year might have its origins here as well - without their partisanship, poll respondents had little to go on except their vague sense of the media's consensus view of the race. Predictably, they claimed to support Clinton. Finally, this might account for momentum. Voters take a close look at winners at precisely the moment they are basking in the glow of positive media coverage. Unsurprisingly, researchers have found that more informed voters are less susceptible to momentum effects.

These considerations have two implications. The first is good news for pollsters: life will get better for you! When we move into the general election - the polls will settle down and start agreeing with one another. I would note the stability in those head-to-head match-ups. They have barely budged an inch even as the race in both parties has been chaotic. Given that voters have their party identification to ground their general election responses to the pollsters - it makes sense they would be steady. The second is a warning to consumers of political news: continue to be wary of these primary polls. Without partisanship anchoring vote choices, they are still prone to large dramatic shifts at the last minute. That's what happens when voters try to make decisions without their use of most trusted cue, their partisanship.

-Jay Cost

The Democratic Race Moving Forward

Today's column is a continuation of yesterday's essay. Thus, it is appropriate to review exactly what was accomplished yesterday. Using two distinct data sets - the exit polls to date supplemented by the statewide vote totals - we formulated a basic outline of the set of groups that seem to be influential in the Democratic primary race.

The exit polling data indicates that several variables are factors in determining who wins a given primary or caucus: union workers, Catholics, Hispanics, African Americans, party identification, geography, and gender. In particular, we noted yesterday that southern whites tend to support Clinton, while northern whites split their support between Clinton and Obama according to gender.

We supplemented this analysis with an OLS regression model whose purpose was fairly modest. It was designed to confirm some of the observations made on the micro level as well as test whether some other factors might be influential. It was helpful in this regard. In particular, we found that caucus states, states with high white median income, and "homogeneously white" states all tend to support Obama.

What I would like to do today is use this knowledge to estimate where the race goes from here.

Hillary Clinton has been on a bit of a losing streak lately - and that streak continued last night. Clinton lost all three contests in the so-called "Potomac Primary" by large margins. Unsurprisingly, there have been stories about the bottom dropping out of her campaign - and we should expect them to continue given last night's results.

However, I respectfully submit that all of this talk is a bit hasty. Not necessarily wrong. Just hasty.

Of course, she has lost the seven contests since Super Tuesday - the District of Columbia, Louisiana, Maine, Maryland, Nebraska, Virginia, and Washington have all gone to Obama. However, if we understand the Democratic race as the mobilization of demographic groups - none of these losses should come as a surprise, including the states she lost by a large margin. Each of the seven post-Super Tuesday states played heavily to at least one of Obama's several strengths.

For instance, the following chart reviews the African American population per state, the median white income per state, and whether the contest was a caucus.

Chart 1 2-13.jpg

These contests are tailor-made for a candidate that fuses the coalitions of Hart and Jackson, and one who inspires tremendous enthusiasm among his supporters. Given the voting coalitions that have formed over the last month and a half, Clinton never really stood a chance in any of them. African Americans drove Obama's victory in Louisiana. In the District of Columbia, Maryland, and Virginia, African Americans combined with wealthy whites to secure him victory. In Maine, Nebraska, and Washington - Obama took advantage of largely homogenous white populations and caucus contests to secure victory.

In other words, it is hard to identify a momentum effect here. Clinton's losses in the contests are as explicable as any of her losses before or on Super Tuesday. Obama has systematically won states that play to his particular strengths since the Iowa caucus. So has Clinton. Her problem has been that she has not had any good states in the last week.

This is not to say that momentum is not playing a role. The point here is more modest: if it is playing a role, it is currently undetectable. Relatedly, this is not to say that momentum will not develop as we move forward. It very well could. If it does, what we will see is Obama doing better among Clinton's strong groups, or him consolidating his position among his strong groups. However, I noted back in December that this the kind of momentum - where a contest in moment one influences a contest in moment two - tends not to describe candidates like Clinton and Obama. Front-running candidates with high name recognition, secure bases of support, and money to spend usually win or lose depending on how much money they spend and whether the state plays to their strengths. The most immediate results tend not to be factors for front-runners (though there are exceptions, e.g. George H.W. Bush in 1988). If front-running candidates benefit from momentum, it is usually via a process similar to what has aided McCain: they are launched by big wins in one of the early states (usually New Hampshire).

What can we expect to happen in the future? It is difficult to say. We can make a few modest statements if we assume that what has generally been true in past past cycles holds true this cycle. Namely, let us assume that momentum does not develop, and that both candidates face a hard slog through the sixteen remaining contests.

The following chart reviews the remaining sixteen states according to several of the variables we outlined yesterday.

Chart 2 2-13.jpg

The final two rows are worth particular attention. These are averages, weighted by total pledged delegates per state, of the African American population, the Hispanic population, median white income, and the union population for contests that have yet to occur (top row) and and those that have already occurred (lower row).

A comparison of these rows gives reason to think that the remaining contests will be as tight as the past ones. Note first of all that the median income of whites drops from here on out. This augurs well for Clinton, who seems to do better among "downscale" white voters - and it stands to reason that as the median income of all white residents in a state declines, so also does the median income of white Democratic voters (this is presumably why this factor was shown to be statistically significant yesterday). More good news for Clinton, though it is not represented in the above chart, is that there are just two caucus states left. Obama, however, is advantaged because the number of African Americans remains roughly constant, and the number of Hispanics and union workers declines. What is more, there are about as many "homogeneously white" states to go as have already occurred.

Examining matters from another direction yields the same basic point. Looking at the above states, we could easily envision Obama doing well in states like Hawaii, Montana, Oregon and a few other smaller ones. Clinton, for her part, should do well in states like Kentucky, Rhode Island, and West Virginia. Texas and Ohio play to many of her strengths, and she should be quite competitive in Pennsylvania. Though she seems down now, and though there are reasons to favor Obama in several of the above states, Clinton has real strengths in several small states and many of the big ones.

All in all, this implies a rough parity from here until the end of the primary season. Approximately speaking, neither candidate seems to have an advantage in the remaining contests. So, my suggestion to readers is not to get caught up in the "Obama is inevitable" storyline. Minimally, we should all remember how well the "Clinton is inevitable" storyline worked out five months ago!

Again, these considerations assume stable voting coalitions, and therefore an absence of momentum. This assumption might not hold. If it does not, what we will see is Clinton start to lose portions of her strongholds, or Obama consolidating support in his. Unfortunately, the exit polls in Virginia and Maryland do not provide much of a clue about whether momentum is coming into play. Obama's best groups are heavily represented in both states - and the exit polls do not really dig deep enough into voter demography to offer a clear answer as to whether Clinton is hemorrhaging parts of her core constituency. For instance, the exit poll shows Obama winning white men in Virginia by 14% and Clinton winning white women by 9%. The white gender gap remains, but it favors Clinton less. In Maryland, the story is the same. White men break more heavily to Obama than they have in the past; white women break less heavily to Clinton. Is this simply a function of wealthy voters, male and female alike, going for Obama? Amazingly, 39% of Virginia Democrats and 41% of Maryland Democrats reported incomes of $100,000 or more - this plays to a major strength of Obama. If income is causing these changes in the gender gap, it is hard to see momentum as a factor. If it is something other than income, Obama might indeed be benefiting from momentum.

-Jay Cost

On the State of the Democratic Race

The Democratic race is as tight as can be. Clinton and Obama currently split the pledged delegates as well as the popular vote. Why is this the case? What types of voters are coalescing around these candidates? Prior to Super Tuesday, I offered several essays on the matter (see here, here, here, here, and here). However, I was not able to offer conclusive statements. There simply had not been enough contests. Thanks to Super Tuesday, matters have changed. It is now possible to submit something more complete.

Since the publication of Berelson, Lazarsfeld, and McPhee's Voting in 1954, the study of groups has been a cornerstone of voting analysis. Voters frequently behave similarly to the way that others with similar relevant characteristics behave. If we know what those characteristics are - we can understand voter behavior better.

Currently, we have two types of data available for understanding group behavior in the Democratic race. We have micro-level survey data that comes to us in the form of the exit polls. We also have macro-level data that comes from the statewide results. We shall use both sets to understand group behavior. This puts us in a good position: because neither set is perfect, one can supplement the other. In particular, the exit poll data is incomplete. The media only releases select results from the exit polls - so that limits its utility. The media also tends not to poll caucus states, which have favored Obama to date. Thus, our analysis of the exit polls will "skew" somewhat toward Clinton. We should account for this when we interpret the data. We should also use the macro-level data to confirm the conclusions we draw from this micro-level set.*

Methodological considerations aside, what can the exit polls tell us about the coalitions of Clinton and Obama?

On the one hand, they confirm much of what we already knew. Consider the following chart, which reviews Obama and Clinton's share of several relevant groups weighted by their share in the population across all the voting states*:

Chart 1.jpg

At first blush, we see the same picture we have seen since New Hampshire. Clinton is winning a "traditional" Democratic voting coalition. It is centered around women and Hispanics - and includes voters with lower incomes, self-identified Democrats, union workers, and Catholics. Obama, for his part, is drawing a combination of the Jesse Jackson and Gary Hart voting coalitions - African Americans, wealthier voters, self-identified Independents, non-union workers, and white men.

However, there are some peculiar features here. In particular, Obama wins African Americans overwhelmingly, yet Clinton wins voters who make less than $50,000. This is noteworthy, given that nationwide African Americans make less than whites. Notice also that he loses white voters even though he wins Independents decisively. Again, Independents tend to be white. This implies that there might be some distinctions that the data as it is presented is simply not picking up.

Accordingly, I reexamined the same demographic groups, this time dividing the data according to southern and non-southern subsets. The results are arrayed in the following table:

Chart 2.jpg

As you can see, splitting the data according to region uncovers some intriguing divergences. Note first that Obama wins all demographic groups except whites in the South. He wins voters regardless of sex, income, party registration, or union affiliation. Why? African Americans. They have comprised about 41% of the southern vote to date - and they have broken for Obama by 69 points. In the South, we can also see the racial gap we saw in the earlier chart grow larger. Clinton wins all white voters by 13%. She wins southern whites by 28%.

Look outside the South - to the two columns on the right. In the non-southern states, we find a fascinating twist. The racial gap transforms into a typical gender gap. Obama wins white men by 8 points. Clinton wins white women by 19 points.

This implies that race is playing a different role depending upon the location of the contest. In the South, there is a racial divide. Clinton wins white voters. Obama wins African American voters. When African Americans make up a strong share of the vote (e.g. Alabama, Georgia, Louisiana, and South Carolina), he wins. When they do not make up a strong share (e.g. Oklahoma and Tennesse), she wins.

In the non-South, matters are more complicated. African Americans still go heavily for Obama - but whites are split. White men prefer Obama, white women prefer Clinton. By itself, this favors Clinton because white women go more strongly for Clinton than white men go for Obama; what is more, white women consistently make up a larger share of the vote. Of course, if the groups among which Obama does well are populous in a given state - he overcomes this gender gap. If Clinton's groups are strong, he doesn't.

Remember that we have excluded many caucus states that broke heavily to Obama. This implies that Obama probably does better than the above results suggest. Where he specifically does better, and how much better he does - we cannot know. However, the macro-level data can offer a way to confirm the demographic trends we have found.

Accordingly, I have run an ordinary least squares (OLS) regression analysis based upon the statewide results. OLS regression is a statistical tool that tests whether an explanatory variable accounts for variation in an dependent variable, controlling for other explanatory variables.* It thus provides a way to determine if certain demographic groups are separately influencing statewide results in a given state. Our dependent variable is the difference between Obama and Clinton's share of the results. We shall test several independent variables:

(1) Median income of whites per state. Obama's strength among non-southern white males, Independents, and higher income voters suggest that white voters break for Clinton or Obama by economic lines. The theory tested here is that as white voters make more money, they become more inclined to vote for Obama.

(2) Whether the state is a primary or caucus. Obama seems to do better in caucuses - perhaps because it takes more dedication to attend a caucus, and his voters seem to be more intense.

(3) Number of candidate visits. This is designed to measure campaign effects per state. It stands to reason that the more candidates visit a state, the more money they have poured into the state. This might influence vote turnout and therefore state results.

(4) Whether a state is "homogeneously" white. This speaks to an intuition I expounded after the South Carolina results. My theory is that white voters in states with large white majorities see Obama as an insurgent, independent-minded candidate. Meanwhile, voters in more diverse states see him differently. It was noted above that Obama picks up pieces of the coalitions won by Jesse Jackson and Gary Hart. Perhaps white voters tend to see Obama as Jackson or Hart depending upon the racial demography of their environment. Accordingly, "homogeneity" is measured via whether Hispanics and African Americans together constitute more or less than 10% of the population.

(5) Whether the state is a Southern state. There seems to be a unique effect among white voters depending upon the region. According to the exit polls, Clinton does better among southern whites than she does among northern whites. This variable will catch distinctions that show up on the macro level.

(6) Percentage of union workers per state.

(7) Percentage of Catholics per state.

(8) Percentage of African Americans per state.

(9) Percentage of Hispanics per state.

Again, these variables were used to predict the difference between Obama and Clinton's returns per state. The results are interesting. The model accounts for 69% of all state-by-state variation between the two candidates [that is the "adjusted r-squared" value]. What is more, eight of the above nine variables were found to be statistically significant at the 95% confidence level (or greater). The only exception was the percentage of Hispanics per state.* This means that we can be confident that all but one of them have influenced the state-by-state results.*

What exactly did we find?

(1) As the median income of white voters increases, Obama does better. This is consistent with the hypothesis offered above: wealthier whites are attracted to Obama, poorer whites are attracted to Clinton.

(2) Obama does better in caucuses than in primaries. This was the strongest predictor of all explanatory variables, which is not surprising in light of Obama's large victories in the caucus states.

(3) Clinton does better as the number of candidate visits increases. This was a bit of a surprise, but it is good news for her. Campaign effects seem to incline the electorate to her.

(4) Obama does better in states that are "homogeneously white." This is consistent with the hypothesis we offered: white voters in "homogeneous" states see Obama differently than white voters in heterogeneous states.

(5) Clinton does "better" as we move to the South. This might sound counter-intuitive. However, remember that we included this variable to account for the inclination of southern whites to go for Clinton. Obama's strength in the south is accounted for by the African American variable.

(6) As the union population increases, Clinton does better.

(7) As the Catholic population increases, Clinton does better.

(8) As the African American population increases, Obama does better.

Regression models like these have two important uses. First, they enable us to predict what will happen in the future. That was not the intention here. The point was not to offer predictions about what will happen next - and rightly so. The model's predictive power (69%) is very high from a certain perspective. From another perspective, though, its accuracy is not great enough to admit of "publishable" predictions - not when candidates are often separated by tiny margins. Tomorrow, I hope to take some tentative steps toward reviewing what to expect in the upcoming contests. The model (refined by today's results) will serve as a foundation for this analysis - but it will not be used for simple divination. It is simply not precise enough.

Another use of regression models is that they isolate and identify influential factors. This model definitely serves this purpose - it confirmed much of what the micro-level analysis showed and it elucidated some new trends. All in all, we have made some important steps - especially when we combine the macro analysis with the micro analysis. We have found that both candidates are putting together diverse voting coalitions that differ according to region. There is evidence that Obama wins Independents, African Americans, white males in the North, "upscale" white voters, and white voters in homogeneously white states. He also seems to do well in caucus states where enthusiasm is a factor. There is evidence that Clinton wins Democrats, Hispanics, white females everywhere, white males in the South, "downscale" white voters, Catholics, and white voters in heterogeneous states. She seems to do better in a state the more attention is paid to it.

As I said, it is not clear which candidate's voting coalition will be larger when all is said and done. Both of them are diverse and quite large. We will talk tomorrow about what to expect from these coalitions moving forward.


[*] Another caveat is appropriate to mention here. We have exit polling data from twenty-three states. Data from eighteen of them is used - excluding Florida and Michigan because the states were not contested. Also excluded are Illinois, New York, and Arkansas because each of them voted heavily for the "favorite son" or "favorite daughter." Our goal is to draw an inference from this data to the rest of the contests. We cannot do that if our data has major exceptions that will not recur between now and the convention.

[*] These figures were computed in the following way. A given demographic group in a state was weighted first by its contribution to that state's electorate, and second by the number of pledged delegates the state has. This method is different from a simple unweighted average of each statewide result, which would diminish the importance of bigger states and enhance the importance of smaller states.

[*] A difficulty here is that several caucus states report state or county delegate results, not raw votes. The model counts these delegate results as though they are raw vote results. This is obviously not ideal - but the inferential damage from this choice seems minimal. It does not appear that there are any differences from caucus-to-caucus, depending upon whether they report delegates or votes. Obama tends to win a large share of both. So, it does not seem to have a particular effect on caucus states - and the difference between caucuses and primaries is captured via the corresponding dummy variable. An alternative approach would be to exclude the delegate-reporting caucus states from the analysis - but this would exclude several observations, thus reducing efficiency. There are often judgment calls like these to make when dealing with non-experimental data - choosing between running a risk of bias or a risk of inefficiency. Since the bias effect seems to be minimal, and the inefficiency seems to be more sizable - it seems best to include the delegate-reporting states.

[*] Why was our Hispanic measure found to be insignificant? One reason might be that there is not a lot of state-to-state variation in the Hispanic population. Some states have large Hispanic populations, but most of them have uniformly low populations. As an explanatory variable stops varying, it becomes harder for OLS regression to pick up on its effect.

[*] One possible objection to this analysis is that previous results influence subsequent results - and therefore this kind of "cross-sectional" investigation is missing a key explanatory variable. This possibility was tested, and it was found to be unlikely. In particular, the outcome of the immediately prior statewide result was temporarily included as an explanatory variable. This factor was found not to have any influence. A test was also conducted to see whether the model does a better job predicting results prior to Super Tuesday versus results afterwards - which might be the case if the early "beauty contests" influenced the states of February 5th. The model seems to predict both types of states equally well. The upshot of this is consistent with the conventional wisdom that the Democratic race has been bereft of momentum to date.

-Jay Cost

The Romney Campaign, RIP

With Mitt Romney's withdrawal from the Republican race yesterday, it is time to take stock of his candidacy.

It is fair to say that Romney was a polarizing candidate. Few candidates rouse such strongly divergent feelings among his fellow partisans. All campaign cycle, my email inbox has been full of people telling me Romney was the GOP's best hope and people telling me he would ruin the party.

We can state this a bit more formally. The following chart details net positive/negative feelings for the GOP candidates among the general electorate, according to the last few Pew polls. It also lists the percentage of people who could rate the candidate.

Favorability 1.jpg
As you can see, Romney's numbers fell as the campaign wore on. Also, note that they went down as the percentage of voters willing to rate him went up. Of course, they did not drop as precipitously as Giuliani's numbers - but unlike Giuliani, Romney was not plagued by personal scandal during the period these polls cover.

The same basic story holds when we look at Romney's numbers among Republican voters:

Favorability 2.jpg

These are probably the more critical numbers to review, even though they exclude self-identified independents who ultimately voted in an open Republican primary. Here we see that Romney lost all the ground he would lose on net favorability among Republicans before the first votes were cast - i.e. before momentum became a factor. This distinguishes him from Giuliani - whose numbers, as I noted last week and as can be seen above, fell before and after the contests began.

If Romney's favorability rating fell first, did this harm him at the ballot box later - when voters started to focus on the race, undecideds started to decide, and losing candidates started to drop out? Unfortunately, we cannot approach this question directly because we just do not have sufficient data for 2008. However, we do have data for years past that provides a clue. Consider John Geer's Nominating Presidents: An Assessment of Voters in Presidential Primaries. Geer's work examines the attitudes of Republican and Democratic voters in Los Angeles, CA and Erie, PA during the 1988 presidential nomination season. He asked voters why they supported a candidate, and these are the results he found:

Justification for Primary Candidate.jpg

Clearly, personality has a significant effect on vote choice. This should not come as a big surprise. Remember that these are primary voters choosing among candidates of the same party. It is surely difficult for them to see significant issue or ideological differences between candidates - and so it makes sense that personality would be a critical factor.

Of course, Pew's "favorability" and Geer's "personality" do not overlap completely - but they are obviously related. Thus, it is reasonable to infer that Romney's dropping favorability rating hurt his vote totals. It surely did not make all of the difference. Romney - like the rest of the field - was hurt by McCain's victory in New Hampshire, which will be remembered as the critical moment of the race. My point is more modest. Favorability did not make all the difference, but it did make some.

All of this might seem strange, given that Romney was viewed quite favorably by the Republican electorate. The critical point is that McCain was viewed much more favorably. That is the key - not Romney's absolute favorability, but his favorability relative to McCain's. Here I would also recall his national unfavorable rating, which includes strongly unfavorable reviews by independents, who were factors in New Hampshire, South Carolina, and some states on Super Tuesday.

How is it that the balance of the general electorate, and a sizable minority of Republican voters, came to dislike Romney? Unfortunately, there is only so much we can say on this front because Pew (and other pollsters for that matter) did not probe why voters did not view him favorably. Some, like Howard Fineman, would argue that it was due to his insincerity. Maybe. Maybe not. The polling data does not provide much of a clue. It is problematic to assume that the average Republican primary voter noticed what a journalist like Howard Fineman noticed, let alone whether he had the same reaction that Fineman had.

I do think we have enough evidence to make a case that Romney's negative (or "contrast") advertising had an effect on his favorability. If we dig a little deeper into the available data - some interesting trends emerge.

First and foremost, it is simply true that Romney's negative campaign against Huckabee and McCain was risky. I discussed the risks of negative campaigning in December. At the time, I noted that the research of Rutgers' Richard Lau and Gerald Pomper shows that going negative is a tricky task. In 2002, they looked at incumbent senators who attacked their opponents, and concluded, "A full accounting of the evidence suggests that, as often as not, attacking the opponent is a counter-productive campaign strategy to follow."

How might these negative attacks have hurt Romney? My sense is that it likely kept him from winning over those who supported McCain or Huckabee. That is, at its most basic level, it backfired; not only did it fail to convince Huckabee or McCain voters to back Romney, it alienated those voters from him. Pew found that Romney's net favorability rating among McCain voters was just +7 in January and +1 in February; among Huckabee voters it was -9 in January and -4 in February. The only candidate who had so much trouble with another candidate's voters is Giuliani, who was not liked by Huckabee voters. This is different - Rudy's divergence on social issues and his scandal-plagued autumn can explain most of that disregard.

I think the fact that Romney was viewed so poorly by McCain and Huckabee voters, but not Giuliani voters is a consequence of his attacks on McCain and Huckabee. I see no other plausible way to explain this pattern - especially in light of the fact that Romney worked for most of 2007 to woo the social conservatives who comprise Huckabee's base. Why else would the first two sets of voters dislike him, but the third like him?

These low ratings may have done electoral damage to Romney. First, consider Huckabee voters. As Huckabee's prospects declined, we would expect some of his supporters to switch to other candidates. Indeed, Huckabee has fallen about four points in the last month. Romney's negative numbers among Huckabee voters implies that if they were going to leave Huckabee, they would probably not go to Romney. Instead, they would probably go to McCain - who enjoys strong positive numbers among Huckabee voters. Minimally, this should dispel the notion that Huckabee's decision to stay in the race hurt Romney. If anything, it probably helped him.

Second, Romney's negativity might have kept him from picking up Giuliani voters during and after Florida. I would note that McCain had a +69 favorability rating among Giuliani supporters in the January Pew poll. If Rudy supporters were predisposed to like McCain - can we expect them to have reacted positively to Romney's attacks on the Arizona senator? Probably not. Of course, McCain and Giuliani share so much common ground that the former would probably have picked up most of the latter's supporters, anyway - but my sense is that Romney's attacks on McCain did not help him with Giuliani supporters who liked both McCain and Romney.

Ultimately, these considerations remain sketchy due to the limitations of the data. Perhaps at some point, an outfit like Annenberg or Pew will release some detailed analysis of how the supporters of various candidates responded to changes in the race. With the data available, the best I can do is argue that the theory of Romney's negative attacks backfiring is intuitively plausible and consistent with what we know. I would like to say more, but am limited by the numbers I have.

Regardless of whether the negative ads had an effect on his favorability rating - it remains true that it was quite low for a candidate hoping to win a nomination. We know from previous cycles that favorability and vote choice move together quite closely. So, we can conclude that Romney's campaign would have been well served by a more careful maintenance of his public image. It should have been more mindful of how voters felt about him, and it should have taken more decisive steps to win not just votes, but also affection, which makes a difference in a primary campaign.

-Jay Cost

On McCain's Voting Coalition

An argument being proffered by Romney supporters is that McCain's victories in the early states have been due to the conservative vote being split among many candidates. By this thinking, McCain would have lost South Carolina and maybe Florida if conservatives had coalesced around a single anti-McCain candidate. Michael Medved commented on this theory yesterday. I would like to toss in my two cents, as this sort of matter is up my alley.

At first blush, this theory might seem compelling. However, if we take a closer look at the exit polls in New Hampshire, South Carolina, and Florida, we can see that this is actually a problematic assertion.

First of all, we have to clear away some of the theoretical underbrush. What we are talking about here is the idea that the electoral results to date display social irrationality. Is this possible? Absolutely.

Suppose that you prefer ice cream over pizza, and pizza over hamburgers. If you were given a choice among all three, you would choose ice cream. If you were given a choice among just ice cream and hamburgers (with pizza excluded), what would you choose? Again, it would be ice cream. The choice between ice cream and hamburgers is independent from your feelings about pizza, which are irrelevant. So, your choice is independent from irrelevant alternatives.

This is a characteristic of individual rationality. Can we expect society as a whole to act rationally in this way? If we all take a vote on ice cream, pizza, and hamburgers - would we prefer ice cream to hamburgers regardless of whether or not pizza was an option? If pizza does change the preference between ice cream and hamburgers, can we say that society has chosen rationally?

These questions are part of a general subject known as "social choice." It has been of interest to smarty pants for centuries, especially French smarty pants. In the 18th century the Marquis de Condorcet and Jean-Charles de Borda argued over how to design voting systems that avoided irrational outcomes like the one highlighted here. Of course, leave it to an American to prove that the French were wasting their time! This is exactly what Kenneth Arrow did in 1951 with his impossibility theorem (a.k.a. Arrow's Theorem). Arrow later won a Nobel Prize in Economics for his achievement.

What Professor Arrow did was devastatingly simple. He identified a handful of basic criteria that we would expect any rational voting system to meet - and proved that no system could meet every criterion every time. We just finished talking about one such criterion - independence from irrelevant alternatives does not always hold when we move from individual choice to collective choice. Depending on how we conduct the vote, society could change its preference from ice cream to hamburgers because pizza was included or excluded as an option.

Romney supporters are using this concept to argue for their candidate. The final result in Florida was McCain > Romney > Giuliani > Huckabee. According to their argument, if Giuliani and Huckabee were not on the ballot, the collective choice between McCain and Romney would have been reversed. Irrelevant alternatives were not independent - as this argument goes. Take the other conservatives off the ballot, and Romney would have defeated McCain.

Let's dig into the exit polls from New Hampshire, South Carolina, and Florida to see exactly whether this claim bears out.

As I argued last week, feelings about the Bush administration have been influential in determining primary vote choices to date. We should note two distinct features about the vote distribution of Bush/anti-Bush voters. First, McCain consistently wins those with negative feelings toward Bush, and consistently loses those with positive feelings toward Bush. Second, most Republican primary voters have positive feelings toward the Bush administration.

So, for instance, in Florida:

- 68% of Republican voters claimed support for the Bush administration; 32% did not.

- McCain won 31% of Bush supporters, 45% of Bush opponents.

- Romney won 35% of Bush supporters, 23% of Bush opponents.

- Giuliani won 16% of Bush supporters, 12% of Bush opponents.

- Huckabee won 15% of Bush supporters, 10% of Bush opponents.

This seems to bode well for the anti-McCain argument. If you fold up the pro-Bush vote into a single candidates in Florida, the pro-Bush candidate might well have beaten McCain. Perhaps McCain has been winning only because the pro-Bush vote has been split among multiple candidate. This would violate the independence from irrelevant alternatives criterion. Taking all but one of the conservatives off the ballot, by this logic, would flip the results in Florida.

This conclusion is hasty. Romney surely won more pro-Bush voters than McCain, but he did not win that many more. McCain, for his part, won many more anti-Bush voters than Romney did. This could make a big difference.

Let's drill this down with some mathematical analysis:

(1) Suppose a two-way Florida race between Romney and McCain in which (following the exit poll) 68% of voters are pro-Bush and 32% are anti-Bush. Also, suppose that the key variable in determining where the voters of the now excluded candidates go is their feeling about Bush.

(2) In the actual race, McCain beat Romney among anti-Bush voters in a multi-candidate field, 45% to 23%. In a hypothetical two-way race, the proportional result would be 66% to 34%, McCain.

(3) In the actual race, Romney beat McCain among pro-Bush voters in a multi-candidate field, 35% to 31%. In a hypothetical two-way hypothetical race, the proportional result would be 53% to 47%, Romney.

(4) In a hypothetical two-way race with proportional distribution of pro/anti-Bush voters, McCain's vote share would be calculated as: 66% X 32% + 47% X 68% = 53%; Romney's vote share would be 100% - 53% = 47%. Thus, McCain would win a hypothetical two-way race determined under these conditions.

The same goes for South Carolina and New Hampshire. I'll spare you the mathematics and just conclude that if you turn those races into two-man contests following the same rules I used with Florida, McCain would still win both states against his strongest opponents.

Perhaps using feelings about Bush is not the best way to re-allocate votes in a two-man race. Perhaps the best way is via ideology. Accordingly, I re-ran these calculations dividing the electorate into four ideological groups: liberal, moderate, somewhat conservative, very conservative. Once again, I folded the actual race into a hypothetical two-way match up, allocating voters by ideology proportional to how they supported McCain and Romney (or, in South Carolina, Huckabee). Once again, I found that McCain wins all three states.

I would also note that if you wanted to reallocate the vote based upon professed second choices - you would find the same result in Florida. Most Huckabee voters there claimed McCain as their second choice. Romney and McCain were essentially tied among Giuliani voters. Unfortunately, the same question was not asked in New Hampshire or South Carolina.

As I hinted above, this does not necessarily mean that the anti-McCain theory is wrong. Take a look at the Florida example. Among pro-Bush Florida voters who actually supported McCain or Romney, 53% went to McCain and 47% went to Romney. So, that was how all pro-Bush voters were split in the hypothetical two-way match up. However, maybe pro-Bush voters would have broken more heavily to Romney. That could alter the results, depending on how they go. If they go by at least 58/42 - Romney would defeat McCain.

While this is possible, it is hard to argue that it is likely. If anything, voters might break more heavily to McCain than my baseline model implies. According to the latest Pew poll, Huckabee voters have a +36% favorable rating of McCain, but a -4% rating of Romney. It is hard to argue that Romney would take more Huckabee voters than McCain, which is actually what the baseline predicts. The same goes for the Giuliani voters - who in the January Pew poll were more favorable to McCain than to Romney. So, if we alter this assumption to reflect the Pew results - we might see McCain's lead grow.

And so, I do not think one can argue that McCain's wins have been dependent on a divided field. Independents, moderates, and Bush disapprovers have certainly formed the core of McCain's voting coalition. However, McCain has done what most winning candidates do: win his base by large margins while stealing plenty of voters from the other guy's base. McCain does not win conservatives or Bush supporters outright - but he has done well enough with them that he could probably win New Hampshire, South Carolina, or Florida in a head-to-head match up.

I think the Romney supporters are on better empirical ground to argue that their candidate's problem has been that Bush supporters and strong conservatives simply have not made up a sufficiently large share of the vote - and that "true" Republicans need to "wake up." Accordingly, their goal should be to turn out more of their voters. Maybe they will be able to do this in California today.

Relatedly, this highlights McCain's potential weakness. While it is difficult argue that a narrower field would turn any of McCain's wins into losses, the fact remains that each race would probably have been tight. Under my assumptions, McCain never would have won more than 54% of the vote - and his victory in South Carolina over Huckabee could have shrunk to a little less than one-half of one percent. So, McCain has been building a voting coalition that can trump the coalitions of other candidates - but not by much. If the anti-McCain forces could have turned out more of their voters in the other states, they might have won.

-Jay Cost