Predicting the Presidential Election

Predicting the Presidential Election

Trevor Thomas

October 26, 2012

I love Electoral College math. I mean, I teach mathematics and I write about politics, so pouring over various Electoral College combinations is right up my alley. Experts all across the country are telling us that this presidential election is coming down to a handful of “battleground” states. Thus, Romney and Obama are spending virtually all of their time and money in these final days before November 6 in states like Florida, Virginia, Ohio, Iowa, Colorado, Nevada, Wisconsin, New Hampshire, and the like.

For most American voters, the memory of the 2000 election, where George W. Bush beat Al Gore in a narrow Electoral victory (271 to 266) while losing the popular vote 48.4% to 47.9%, is still fresh. The weeks-long battle to count and recount votes in Florida, where the phrase “hanging chads” entered our vernacular, is a path to which no one wants to return. For 36 days, the winner of the 2000 presidential election was in limbo. After countless hours of elections officials leering at ballots, nonstop media coverage, and 47 lawsuits, ultimately it took a ruling by the U.S. Supreme Court to put the matter to rest.

Do not let your hearts be troubled. In spite of the seeming closeness of this race, another outcome like 2000 is highly unlikely. In fact, only four times in our nation’s history has the winner of the popular vote not gone on to occupy 1600 Pennsylvania Avenue.

Prior to the 2000 election, the last time such an event occurred was 112 years earlier, in 1888. Incumbent Grover Cleveland narrowly won the popular vote (eight-tenths of a percent), but Benjamin Harrison easily (by 65 votes) carried the Electoral College. Two other times during the 19^th century the winner of the popular vote failed to win the presidency.

The closest Electoral College result in American history occurred in 1876 when Democrat Samuel Tilden won the popular vote by three percent but lost the Electoral College vote to Republican Rutherford Hayes 185 to 184. Probably the strangest presidential election result was one of the earliest. (An excellent source for such data is Dave Leip’s (online) Atlas of U.S. Presidential Elections.)

In 1824, four candidates received significant support: Henry Clay, William Crawford, Andrew Jackson, and John Quincy Adams. Jackson won the popular vote (41%), followed by Adams (31%), Crawford (16%), and Clay (14%). Jackson also won the Electoral Vote with 99 votes. Adams received 84 while Crawford and Clay received 41 and 37 respectively.

However, Jackson’s Electoral support was not enough to win the presidency. According to the 12^th Amendment, the election then went to the House of Representatives where Adams was the winner.

A significant mathematical note here is that in only one case where the winner of the popular vote lost the Electoral College did the candidate actually receive over 50% of the popular vote: Tilden in 1876 with 51%—and remember this was the closest Electoral result in U.S. history. So out of the 56 U.S. presidential elections, only once did a candidate receive over 50% of the popular vote and not make it to the White House.

All of this is to say that, in spite of the ENORMOUS amount of attention (not to mention polling) paid to a handful of states, perhaps the best indicator of who will win the presidency are the national popular vote polls. If it looks likely that a candidate is going to receive at least 51% of the national popular vote, an Electoral victory is almost certain.

Political expert Charlie Cook (The Cook Political Report) said as much back in June of this year. “All of this time and effort spent parsing state-level polls would be better spent more closely examining the national polling data, particularly looking at how the candidates are performing now compared with Obama and John McCain in 2008, and examining how likely the members of specific (and potentially decisive) demographic groups are to actually vote.”

Cook also notes that, “If a race is close nationally, it will be close in a lot of individual states, too.” His implication above is that the inverse is also true. If the race is not close nationally, then it will not be close in very many states, including the “battleground” states. In other words, once a candidate reaches a particular level of support nationally, any of the states that were particularly close are almost certainly in line with the national vote.

As I indicated above, that level of support seems to be around 52%. If Romney or Obama gets to this number, any states that are seemingly tied are virtually guaranteed to be in the camp of the leader. Therefore, as we approach November 6, keep an eye on the national polls. (Real Clear Politics (RCP) is an excellent source.)

Of course, currently Romney is the candidate in the best position for such an outcome. As of this writing, Romney’s RCP national polling average stands at 48% while Obama is at 47%. Also, Romney has been at or above 50% in about half of the national polls in the last ten days. Gallup, in its six day polling average, has had Romney at or above 50% since October 15. During this same period Obama has been at 47% or below. No presidential candidate that has polled at 50% or better in the Gallup survey by the middle of October has gone on to lose the election. (Gallup has correctly predicted 16 of 19 presidential races dating back to 1936.)

Keep a particularly close eye on the national polls on the Saturday through Monday prior to Election Day. Most polling agencies will then produce their final predictions. These are the polls upon which their reputations mostly rest, and accuracy will be essential. Plus, extreme efforts to predict any voters left undecided (most of which will prefer the challenger to the incumbent) will be taken. Of course, when all else fails, just check in Wednesday, November 7. This (usually) will tell us everything.

At the Intersection of Politics, Science, Faith, and Reason