Wednesday, September 10, 2008

Prez Polls: the economist's view

I had fun wearing my academic hat on the VP pic, so imagine my fun when I looked at the RealClearPolitics poll average lines for McCain vs. Obama with the same eye.

A little 'splaining to do. We are going to do a little "chartist" work and treat the poll averages below like they are two stocks; though because 1v1 polls are a zero sum game the comparison is inexact - the behavior of people is the same.

As a devotee of O'Neil and Kaltbaum, I am a big fan of the "cup with handle," 10-week line, and standard price-volume metrics well known to those who have read William J. O'Neil's, How to Make Money in Stocks. However, there are some that work momentum which is a different discipline altogether.

For polls, a cup-with-handle just doesn't work, and pure momentum doesn't work well in a zero sum game. However, there is one pattern that does work for polls if you have a basic understanding of charts; high and low brackets and tracking the trends of "higher highs" and "higher lows" vs. "lower highs" and "lower lows." If you are lost, read the book - or just stick with me.

Look at the crayons I put on the graph from Tuesday in RealClearPolitics, and let me show you what I am looking at -- and what you should be looking at as we go forward.


First let's look at Obama's chart. For everything, ignore all to the left of March. Just doesn't matter. Note how Obama is setting up a nice climbing trading range, breaking away from McCain in late APR - early MAY - until the first week of JUL. Then he just stalled and then broke through his bottom support line a week later. That is when he started having problems and when McCain started getting his footing. The trend started six weeks prior to Palinomania; the second trend line down. Also of note is the sugar-high spike at the Dem convention. No sustainability of momentum. If Obama breaks and stays below 46 for any length of time - he is in very big trouble.

Now McCain. Just the opposite story - except that a week after Obama started a new downward pattern, his started up, breaking through Obama last week. The questions is - will the Palinomania be a sugar high spike - or the start of a new, higher base?

That is where we are - at a base. A few things are going on here.

-- One line I hear from Dems is that McCain's gain is just Republicans coming home. That is part of it, but that doesn't explain Obama's decline starting in the first half of JUL. Obama was losing supporters. Were they going to McCain, or to the "neutral" zone? Hard to say - but the net loss is clear. This wasn't Palin's fault.

-- After the JUL 4th weekend, McCain started gaining support. Why? What happened? Did he do anything to excite the base? No, he didn't do that until the mid-AUG Saddleback discussions. Could it be the result of the "...awww WTF - McCain..." feel from many Republicans when they saw the possibility of the most Leftwing Senator in the White House? Probably. What we may have seen is a little of a shuffle from "neutral" to McCain as some of the iffy Obama supports went from Obama to "neutral."

Where does that leave us? Well, go back to my comment about forming a new base. Unless one of the four P-VP candidates has a major scandal (the Obama Smearborne invasion of Alaska is begging for that), the next few weeks are going to tell us where we stand in the sprint to NOV.

If you note, I circled 46. Here is why. McCain will come down and Obama will come up as more and more undecided come off the sidelines and the Democrat smears strip off a few weak sisters for McCain-Palin. I believe that Bill Clinton's 1992 43% is the absolute bottom for a Democrat after 8 years of an unpopular Republican. Obama and/or Biden could be found in bed with a dead girl or a live boy and still get 43%. He gets an extra 3% from that fact that Obama has such a sound goodwill among the young and misty-eyed.

McCain-Palin can count on a much lower 33% unshakable base, 4% lower than Bush 41's take in '92. Maverick McCain is getting a 13% bump up via 7% who are McCain Democrats and Independents, and 7% of Republicans and Libertarian Independents who were sobered up by Palinomania.

That gets us to what I think will settle out by 01 OCT to a 46-46 tie, on average - bouncing around the margin of error. Here is where the debates and the big "Mo" comes in. If the highs and lows are trending a little one way or another and the debates are a draw, McCain could have a chance to win. I never thought that past 51% chance until now. He has the momentum. There is not Ross Perot. The Libertarians continue to nominate clowns (remember Bob Barr from Borat?), and McKinney and Nader cancel him out.

Let's give 2% to each side. That leaves 6% to fight over --- that squishy middle. In that case, it is a long election night and all the Electoral College.

That is one option. The other is that McCain keep getting a percentage point or less a week as people come home and others realize what Obama really is. That is a situation where Obama falls below or stalls at 46%. Obama has to stay above 46% in order to stop the bleeding of support. If he can stay at 48%, I think he will win.

The other is that McCain and Palin just plain fail in the debates, then they fade at the end. Then again, expect this to be something nasty as nasty can be. Expect everything to be racist (case in point, Community Organizer is now a CMEO complaint). McCain-Palin get spooked, lose their Mo and fade. They get too careful (like Obama thought he did with Biden) and fade. Your guess is as good as mine - though I never thought it would be like this in the second week of SEP.

I know the ultimate answer is "42," but for this case, it is 46.


Speaking of charts, Dem. strategists Peter Feld has a nice overview here.

No comments: