EDITOR’S NOTE: Using quantitative and other techniques, Reid Holloway published studies in 1996, 2000, 2004, and 2008 forecasting the outcome of the Electoral College. His model accurately forecasted not only the Electoral College outcome but also the popular vote totals for each of the candidates in all 50 states and Washington, DC, within one half of one percentage point of the actual shares. He called 47 states and the District correctly in 2004 and 2008.
Before diving into the data — and please note we’re using date from May 21 — it’s important to note that what I do in predicting presidential elections stands in contrast to the so-called "science" of nose counting, polls, and the assumption that political issues drive votes and voter shifts.
My basic premise is that the issues of the day are really secondary and fleeting and that ongoing dynamic equilibria in demographics are far more important. Every day World War II veterans and baby boomers meet their collective demise and are replaced by voters from Generations X, Y, and so on.
In my view, it is this generational shift that affects the presidential election results far more than debates over gay marriage and other day-to-day issues.
A supporting element to this basic premise on my part is that a presidential election in the United States consists of 51 separate and simultaneous elections. This means the national polls (which are based on the myth that we have one national election) are only indirectly useful by definition until they focus on specific states.
Many may disagree with these overall premises, and I recognize that they entail plenty of risk (particularly when issued this far ahead of the election).
However, I’ll present my findings and you can decide for yourself.
Based on the assumptions that Mitt Romney can't carry any state John Kerry managed to carry in 2004, and that Barack Obama can't carry any state John McCain managed to carry in 2008, I conclude that Obama can confidently count 246 electoral votes in his column, while Romney can confidently count 180 electoral votes in his column.
That's 426 of the total 538 electoral votes and accounts for 41 states and the District of Columbia; 270 electoral votes are needed to win.
Of the remaining 112 electoral votes in nine swing states not satisfied by the "Kerry/McCain" test, six (Florida, Iowa, Nevada, New Mexico, North Carolina, and Ohio) go to Obama (totaling 79 electoral votes) and three (Colorado, Indiana, and Virginia) go to Romney (totaling 33 electoral votes) based on the presidential election histories in each of these nine swing states for 1972, 1976, 1980, 1984, 1988, 1992, 1996, and 2000.
All tallied, with the swing-state results included, this brings the Obama electoral vote total to 325 and the Romney electoral vote total to 213.
I don't view Pennsylvania as a swing state; it's blue in my opinion. Of further note, Arizona is now considered a swing state by many pollsters, and this is McCain's home state. I thought it was iffy for McCain in 2008 but he carried it.
As for Michigan and Massachusetts, I think Romney considers Michigan his home state. This is where he spent a good deal of his youth; his father, the late Michigan Governor George Romney, was an automotive executive turned politician who once headed the now defunct American Motors, engineering a significant turnaround. But there are various reasons for Romney to call California home (the modern-day Romneys own a home there), as well as Utah (the “Mormon state”), and Massachusetts (where the younger Romney served as governor) his home state. By my reckoning, of these states, he only gets Utah.
Is Illinois a problem for Obama? Was Massachusetts a problem for Kerry in 2004? When was the last time, if ever, a president was elected who didn't carry his home state? It’s only happened twice, according to Geoff Skelley, Media Director for noted political scientist and Ph.D. Larry Sabato, who heads the University of Virginia’s Center for Politics: “In 1916, Woodrow Wilson failed to carry New Jersey, the state he governed prior to the presidency, but still managed to defeat Charles Evans Hughes; and in 1844, when James K. Polk lost Tennessee but still defeated Henry Clay.
Of course, identifying a president’s home state can be awkward; Richard Nixon was technically running from New York in 1968 and 1972 but the vast majority of the voters viewed his home state as California considering he was an elected official from there.”
Based on Real Clear Politics
polls, here's the national popular vote round up. Again, note that a presidential election in the US is 51 simultaneous but separate elections in the 50 states and the District of Columbia. Nevertheless, over several months the popular vote polling gap between Obama and Romney has held within a somewhat varying but nonetheless steady range.
What I've done with the Real Clear Politics data is a simple statistical analysis. In three out of three simple tests, Obama remains comfortably ahead: by mean, by median, and by outliers, it all points to an Obama victory.
Even Scott Rasmussen, political analyst, founder and president of Rasmussen Reports, is giving Obama a three-point lead. I highlight Rasmussen’s work because he makes a point (which I consider valid and important) of using voice-unbiased machine-calling "likely-voter"-qualified samples. Many of the other polls on the above list are samplings of "registered voters."
The data that follow recaps my forecast described above, followed by the actual 2004 and 2008 electoral vote results forming the database for the "Kerry/McCain" test taken from the official federal archives.
This go-around is my fifth such study since I started doing these forecasts in 1996. The present forecast:
Results for the 2004 presidential election from the official archives:
Results for the 2008 presidential election from the official archives:
Here's how actual popular vote results on November 4, 2008, compared with the model's forecast four months earlier in July:
Here’s how my current forecast would be altered if Romney carried Florida, North Carolina, and Ohio (along with Virginia, which I’m already giving to him):
This represents a real squeaker, indeed an epic one, if this configuration were to turn into reality. I should again emphasize that I consider this configuration quite unlikely. Then again, anything's possible. And I should also mention that such a scenario would highlight the peculiar nature (and thus exacerbated importance in such a scenario) of the split votes in Maine and Nebraska, which I have treated in a "winner take all" fashion above.