Wednesday, April 23, 2008

Do Campaigns Matter? A Reflection on the Results in Pennsylvania

As I awoke this morning to find the final margin in yesterday's Pennsylvania primary, I was haunted, to some extent, by the parallels that were drawn between Ohio and Pennsylvania in the wake of the Ohio primary some six weeks ago. If you operate under the assumption that the demographics in each state are fairly highly correlated, the results of each primary speak for themselves. In both Ohio and Pennsylvania, Clinton's margin of victory was essentially 10 percentage points. So after all that spending and all the bickering and all the revelations (Rev. Wright, Bosnia, lapel pins, etc.) the result of a six week campaign was essentially nothing. Well, that's the way it looks; cynical as that may seem. [Tom Holbrook and Jim Campbell may want to weigh in now on that question.] I don't really subscribe to that because Clinton's original Pennsylvania poll numbers following the Texas-Ohio results got a boost based on her wins. Over time, though, those numbers decayed and came back down to earth. There was some variation in there based on spending, advertising and other revelations, but in the end the six week long efforts by Obama and Clinton (endogenously) and outside factors canceled each other out.

So do campaigns matter? There is the argument that campaigns cancel each other out, sure, but in the end we only get to see the fruits of a campaign's labor when an election is close. And that's what we have here: a closely contested race for the Democratic nomination. Do campaigns matter? No, if all you're doing is looking at margins between certain states. Yes, if you look at Obama's strength in caucuses or Clinton's approach to Texas and Ohio. Each either knew and exploited the rules or took advantage of some last minute doubt raising.

Of course, if you lean on state demographics as the major indicator of success, then most of the remaining states, save Indiana, fall squarely in either Clinton's or Obama's camps. The real battle then, will be waged there and among, ahem, the superdelegates. The party, I'm sure, is really going to step up the pressure on the superdelegates to decide, one way or the other, sooner rather than later.

And what of momentum? Rob has weighed in in the comments to yesterday's post. Is it dead or is it just the uniqueness of this Democratic race that has made it a non-factor? The comments await.

Who's next? Well, Guam is officially up next followed by North Carolina and Indiana. Guam is Obama country and North Carolina appears to be as well. Let's see how yesterday's results get spun though and how the polls move in the meantime. Two weeks is shorter than six, so Obama doesn't have as much time to get things back to "normal" after what should be something of a Clinton boost after yesterday. But then, momentum may not play a role at all.

The Pennsylvania aftermath has pushed the 2008 Electoral College Maps back a day, so I'll be back with those tomorrow.

12 comments:

  1. Josh, to test the idea that campaigns do not matter, we would need one side to serve as the control group, to not campaign. Do you think that you can talk the Clinton or Obama campaign into postponing efforts for the sake of social science and to put that null hypothesis to the test?

    ReplyDelete
  2. Rich,
    I've been in negotiations with the Clintons about that, and they were close to caving to my demands (in the name of science). They got back to me late last night, though, and told me that the voters of Pennsylvania were not agreeable to the terms the Clintons and I had worked out prior to yesterday's contest.

    Oh well. Back to square one.

    It's funny: Mike Huckabee has gladly agreed to stop campaigning (though not in the name of science) to help me out.

    ReplyDelete
  3. Josh,

    Interesting perspective. I realize that you are posing a spurious question to get a response. If a candidate does not campaign at all, it sends a signal of lack of interest or a lack of resources. All but the most dedicated voters are then going to abandon you. When a campaign is engaged, however, then the candidate is in play. This year among the Democrats when both candidates contest the race, demographics appear to take precedence over momentum. Much has been made in the last 14 hours of Obama not being able to close the deal, particularly in NH and on Tsunami Tuesday. However, if Bill had not played the race card (or fallen into the nefarious trap set for him by the Obama campaign) or if Hillary had contested half the states in the winning streak after Tsunami Tuesday, she'd be the one calling for Barack to get out of the race for the good of the party.
    From a momentum perspective, a look at the Zogby tracking polls there did seem to be some significant changes:

    4/9-10 Clinton +4

    4/11 Bittergate

    4/15-16 Clinton +1

    4/16 Philadephia debate

    4/16-17 Clinton +3
    4/17-18 Clinton +5
    4/18-19 Clinton +3
    4/19-20 Clinton +6
    4/20-21 Clinton +10

    These figures from Real Clear Politics, if they represent real trends, suggest that Bittergate and the debate let Clinton back into the race and that the negativity favored Hillary over Barack.

    So, where do we go from here? Obama needs a win of 10% or higher in NC and a win or close loss (less than 5%) in IN. (Sorry Guam, I don't think many superdelegates will be paying much attention to you!) If he achieves both of these feats, I think you will see superdelegates start flocking to Obama. On the other hand if she loses NC by less than 5% and wins IN by more than 5% , he is in serious trouble and the delegates from FL and MI begin to come into play and a credential fight is almost a certainty.

    ReplyDelete
  4. Rob,
    That lack of interest/lack of resources candidate sounds an awful lot like Fred Thompson actually. I knew Huckabee volunteered too quickly. I should have gone after Thompson.

    I agree on the numbers you settled on for IN and NC. Those are big ifs though. Ifs that got smaller last night. 100 may be the magic number after all.

    Plus, Jesse Helms seems to have gotten involved in the NC race. Not really, but the GOP in NC has prepared and is planning on running a Wright ad attacking both Dem. gov. candidates who have endorsed Obama. Incite those racial tensions in the same way Helms' "white hands" ad stirred it up in his 1990 senate race against Harvey Gantt. McCain has already denounced the ad.

    ReplyDelete
  5. Oh, I should add that I miss NC politics. As a native, it is killing me that I'm missing a competitive presidential primary there.

    ReplyDelete
  6. Thanks. It will be interesting to see if Hillary denounces it as well. Betcha she does not unless and until she gets considerable heat for remaining silent. As I recall, Helms needed to get (and did get) over 70 or 75% of the white vote to win that race. I remember Harvey Gantt well as the first African-American to enroll in Clemson. I was in junior high at the time living in Clemson, and we were told to stay off the streets. He was a class act and was readily accepted at a very conservative school. The major regret was that he didn't try out for the football team as he had played as a runing back at Iowa State if memory serves me correctly.

    See you soon.

    ReplyDelete
  7. Looks like the media finally figured it out!

    http://politicalticker.blogs.cnn.com
    /2008/04/24/was-pennsylvania-a-double
    -digit-clinton-win-or-not/

    ReplyDelete
  8. I notice that Clinton is now claiming the higher popular vote when you count in MI, she is ahead by 122,471 votes. It also does not include votes from IA, NV, ME & WA. Her claim might be a tactical mistake. If Obama can come back now and overcome the margin using her rules, she has undercut her own argument. Of course she doesn't seem to be holding to the magic number of 100 delegates mentioned in one of your earlier entries.

    ReplyDelete
  9. Rob,
    I read somewhere today that Clinton's estimate included "estimates" of caucus participation in those states. Ah, here's the link. Yeah, moving the goalposts (again) could hurt her, but her campaign has done a magnificent job doing it so many times already (without terribly too much scrutiny). I don't know that many of the superdelegates are buying it. There is at least one who is willing to wait and see if she makes up ground via her new yardstick.

    ReplyDelete
  10. According to RCP, with those caucus votes she is ahead by 12,249 votes. That margin is unlikely to hold up. CNN seems to be holding to the 500,610 number that has Obama in the lead. I t does not count FL MI or those four caucus states that have not released the popular vote totals.

    ReplyDelete
  11. I wonder if the Clintons' accountant is as good at fudging numbers as her delegate and popular vote counters are. Maybe those tax records need to be reexamined.

    I may even have to make that its own mock post.

    Gee, maybe I'll actually have some posts this weekend.

    ReplyDelete