Saturday, September 6, 2014

An Older Prediction - Time to Get your Stats On

Taken from a Facebook note I wrote & posted on November 2, 2012.  Such a simpler  with an update in the post-script!

The 2012 Presidential Election: My Prediction

November 2, 2012 at 3:21pm

Election predictions, essentially, are a parlor game.  That said, if you are in the campaign industry, are fortunate enough to get it right, and enough people know that you got it right, you can wear that laurel wreath as a symbol of your political genius for a decade. (For those in the industry, you know who I am talking names please).

In some election cycles, calling winners and losers at this stage of the game, the weekend before Election Day, is relatively easy.  At this point, former Vice President Mondale knew he wasn’t going to defeat President Reagan.  Senator McGovern knew that he wasn’t going to deny President Nixon a second term.  I am quite certain that Judge Parker looked at the map when he was running against President Theodore Roosevelt and said the 1904 equivalent of “Gentlemen, I am hosed.”

In other cycles, including some recent presidential elections, the outcome was far more uncertain.  Both 2000 and 2004 fell into that category.

Bear in mind, a national presidential election is really 51 state elections (or 50 states + DC if you want to get technical about it).  Electoral college votes determine the winner of the election and they are awarded on a winner-take-all basis.  Win a state by a million votes?  Congratulations, you receive 100% of the electoral votes for that state (we can discuss Maine and Nebraska later, I am making a point here).  Defeated in a state by one vote?  Too bad, because the winner just won 100% of the electoral votes for that state. Loser.

In short, small differences in one or a handful of states can determine the outcome.

Everyone knows that a flip of just 269 votes in Florida would have tipped the election from Governor Bush to Vice President Gore in 2000.  Common knowledge.

Fewer people know just how close 2004 was.  To this day, I am convinced that Senator Kerry’s flip-flop gaffe cost him the election.  Some people know that if 59,300 votes in Ohio flipped from President Bush to Kerry (only one percent of the state vote and .048% of the national vote), then John Kerry would have been our nation’s 44th President.  Fewer still are aware that a flip of even fewer votes (18,776 or .015359% of the national vote) would have sent the election to the House of Representatives.  This would have been accomplished by the following flips in three states:

Nevada 10,751 votes from Bush to Kerry (5 electoral votes)
Iowa 5,030 votes from Bush to Kerry (7 electoral votes)
New Mexico: 2,995 votes from Bush to Kerry (5 electoral votes)

So by a relatively small percentage of the vote in three states, 17 electoral votes would have gone from Republican to Democratic, resulting in a 269 – 269 tie and a nation full of Constitutional scholars debating “Does our Presidential Selection Process Work?”  (Note: this assumes that the Democratic elector who voted for Edwards, and I wonder how that Minnesota elector feels about that today, would have cast their ballot for Kerry instead of playing around).

In any event, 2004, based solely on the criteria of the percentage of the national vote that, if flipped, would have changed the outcome, was the sixth closest presidential election ever.  It was even closer than the fabled Truman – Dewey showdown of 1948.

Percentage of the National Vote Difference (flip from eventual winner to eventual loser) from smallest to largest percentage for the seven closest presidential elections:

1. 2000 (Bush vs. Gore) - .00025% (269 votes flip in Florida & Gore wins)
2. 1884 (Cleveland vs. Blaine) - .0052% (524 votes flip in New York & Blaine wins)
3. 1876 (Hayes vs. Tilden) - .00529% (445 votes flip in South Carolina & Tilden wins)
4. 1916 (Wilson vs. Hughes) - .00923% (1,711 votes flip in California & Hughes wins)
5. 1960 (Kennedy vs. Nixon) - .01052% (7,245 votes flip in four states to Nixon & the election goes to the House) or .01777% (12,236 votes flip in five states & Nixon wins outright).
6. 2004 (Bush vs. Kerry) - .01535% (18,776 votes flip in three states & the election goes to the House) or .04848% (59,300 votes flip in Ohio & Kerry wins outright).
7. 1948 (Truman vs. Dewey) - .02564% (12,487 votes flip in two states & the election goes to the House) or .06016% (29,294 votes flip in three states & Dewey wins outright).

Yes, I spent part of a vacation in 2010 working out the math for every presidential election. Everyone has hobbies...

I know, enough history already. So, let’s turn to 2012.

First, I believe that both President Obama and former Governor Romney start off with a base of around 125 electoral votes.  Remember, it takes 270 electoral votes to win.

If either Obama or Romney were to lose one of the states in their “core” (for example, if Obama lost California or if Romney was defeated in Texas), then do not expect a long night, as a landslide is about to occur.  In which case, many pollsters should be drummed out of the industry.  This scenario is unlikely to happen.

There is a second tier of states for each candidate where an upset is very unlikely, but possible.  For Obama, this would be a state such as New Mexico.  For Romney, it would be South Carolina.  Adding such states up with the mortal-lock states, you end up with around 180 electoral votes apiece.

There is a third tier, which is small for both Romney and Obama, where they have a lead but could lose.  Minnesota & Oregon for Obama and Arizona for Romney fit this bill.  If you add up the electoral votes in this category, the President is right around 200 electoral votes and Romney is slightly above 190.  Overall, it is still very close.

This race fundamentally comes down to 11 (and most likely nine) states: Ohio, Nevada, New Hampshire, Virginia, Florida, North Carolina, Wisconsin, Colorado, Iowa, Pennsylvania and Michigan.  These states, primarily the first nine, will play an out-sized role in determining if Obama will be re-elected or if Romney will become the 45th President.

I expect that certain states that voted for Mr. Obama in 2008 will end up in the Republican column in 2012.  I do not believe the President will win any states that voted for McCain in 2008.   That, combined with reapportionment, gives Mr. Romney a much better shot at 270 compared to the previous Republican presidential nominee.

That said, I expect that President Obama will win both Pennsylvania and Michigan, putting him 237 electoral votes, 33 short of victory.

I predict that the President will win in Ohio, Nevada, New Hampshire and Virginia.  Moreover, I predict that Romney is likely to carry Florida, North Carolina, Wisconsin, Colorado and Iowa.

Turning to the two states that factor in performance at the congressional district level when awarding electoral votes, assuming that both Maine Congressional districts vote for Obama and that Romney wins all of the Congressional districts in Nebraska, that works out to 278 electoral votes for President Obama and re-election…and 260 electoral votes for Mitt Romney and a return to the private sector.

More important than the prediction is the possibility that only a small number of votes will determine who wins the Presidency.  In short, vote.


So yes, I called it conservatively for President Obama - who captured 332 electoral votes compared to 206 for Romney, definitely out-performing my projected outcome.  Call me a pessimist.

But with a flip 214,764 votes in four states...Florida, Ohio, Virginia and New Hampshire, Romney would have obtained 270 EVs, just enough to win.  That number of popular votes may sound like a sizable number, and it is...but it represents a mere .166% of the popular votes cast for President that year.  Thus, it was the 15th closest Presidential election in U.S. history, based on my "vote flip" model.  In short, looking at relatively recent presidential election cycles, it was closer than the 1992 election (19th overall) and not as close as the 1976 election (11th overall).

Just some fun facts on a Saturday afternoon.  Not that I am thinking about 2016.  Not at all.

Stay tuned, as more will follow.


No comments:

Post a Comment