Romney or Obama? Political scientists make their predictions
By Dan Balz,
Are you ready to call the election? Mitt Romney certainly isn’t, nor for that matter is President Obama. But a few hardy academics have done so. Out now are a baker’s dozen forecasts produced by political scientists that predict the outcome in November.
Polls give Obama the advantage, nationally and in most of the battleground states, but they are, as is often said, snapshots in time, not predictions of the future. The election forecasts are in fact predictions, based on various and varied statistical models. Most give the advantage to the president, but the verdict is not unanimous.
The 13 projections are contained in the new issue of PS: Political Science and Politics, which is published by the American Political Science Association. Eight of them project that Obama will win the popular vote; five say the popular vote will go to Romney. But the degree of certainty in those forecasts differs. One projection favoring the president says there is an 88 percent certainty that he’ll win, while two others forecasting Obama say there is only a 57 percent certainty.
James E. Campbell, the department chairman at the University at Buffalo in New York, who wrote the introduction to the package, rates them this way: Five predict that Obama will win a plurality of the two-party vote, although three are on “the cusp of a toss-up.” Five predict that Romney will win the plurality of the two-party vote. Three are in what he calls the toss-up range.
One of the most bullish of the Obama-will-win projections comes from Helmut Norpoth, a professor at Stony Brook University, and Michael Bednarczuk, a grad student at the University of Wisconsin-Milwaukee. They wrote that Obama will defeat Romney “by a comfortable margin.”
Their projection, made 299 days before the election, is based on a model that takes into account the performance of the candidates in the primaries and presidential election cycles. “In plain English,” they wrote, “Obama has history on his side as well as the fact that he was unchallenged in the primaries.”
One of the most bearish about the president’s prospects is Alfred G. Cuzan, the department chairman at the University of West Florida. He notes that since 1880, a sitting president has lost his reelection bid only six times, and only twice when the incumbent had succeeded a president of a different party.
But Cuzan, whose model is called the “Fiscal Model,” looks at changes in government spending relative to the size of the economy as his guide. He argues that the expansionary spending policies of the president dim his chances of winning.
“Even if he does squeeze by the Republican candidate,” Cuzan wrote, “it is highly likely that President Obama would do so with a smaller share of the vote than in 2008, the first president in well over a century to be reelected to a second term by a thinner margin of victory than he received the first time around.”
Alan Abramowitz, a professor at Emory University, looks at the advantages of incumbency, presidential approval as of the end of June in an election year and change in real gross domestic product in the second quarter of the year. He calls his method the “Time for a Change” model. He also has made adjustments to factor in the increased polarization in the electorate, which he says has affected the impact of certain fundamentals that generally determine the outcome.
He projected a one-point margin for Obama in the popular vote, but added: “Barring any changes in the second quarter GDP estimate, this is the closest popular vote margin predicted by the model in the entire postwar era although it is only slightly smaller than the 1.2 point margin predicted for Jimmy Carter in 1976.”
That was before the government revised its second-quarter real GDP growth estimate down from 1.7 percent to 1.3 percent. I e-mailed Abramowitz to ask whether he would change his forecast. He said the idea is to forecast the result before the conventions. He said the new GDP number would reduce Obama’s predicted margin but said the president’s better approval ratings would somewhat offset that. He also said he believes the race will tighten before November.
Campbell also foresees a close outcome, but he still tipped in Obama’s direction. He uses two different models, although both include real GDP as one of the factors. Incumbency, he notes, is one big advantage for the president, while the economy is obviously Obama’s problem. In terms of economic growth, he noted, Obama ranks eighth out of the past 10 presidents who sought reelection.
Veteran modeler Michael Lewis-Beck of the University of Iowa and Charles Tien, the department chairman at Hunter College in New York, offer contrasting forecasts based on competing models.
A traditional “Jobs Model” shows Obama in deep trouble. But using a different model, they see Obama winning. Forced to choose between the two, they stick to the jobs model, which shows Obama capturing about 48 percent of the vote. Conceding that any inherent margin for error could result in an Obama victory, they nonetheless concluded, “It still suggests an Obama victory is unlikely.”
Robert Erikson, a Columbia University professor, and Christopher Wlezien of Temple University use a wide variety of economic measures in their “Leading Economic Indicators and the Polls” model. They noted the disparity between perceptions of business conditions and leading economic indicators on the one hand and income growth on the other. Four years ago, the first two were at historic lows while income growth was “middling.” This year it is the reverse.
They ask: “What does this suggest about President Obama’s electoral fate? Is it a dismal election-year economy that dooms the president to certain defeat? Or are economic circumstances brighter than the income numbers would indicate, offering promise of reelection?”
Their answer, made a month before the party conventions, was for a very close election with Obama slightly favored, despite the fact that economic conditions alone would seem to make Romney the heavy favorite.
Douglas A. Hibbs, a retired professor of economics and political science, uses a “Bread and Peace” model — the classic construct of peace and prosperity as the only two factors that really count in presidential elections. He looks at per capita disposable income and U.S. military fatalities in foreign conflicts.
Hibbs projects that, on the basis of his model, Obama will almost certainly lose to Romney. But he is quick to note that he “deviates substantially from prevailing views — particularly those in which I generally put greatest stock: betting price data at the Iowa Electronic Market and Intrade,” which he notes has been bullish about Obama’s chances of winning.
Thomas Holbrook, department chairman at the University of Wisconsin-Milwaukee, uses only a few measures in his model: presidential approval and satisfaction with personal finances, while taking into account whether one candidate is an incumbent. “This portends a close election, but one in which Mitt Romney is the favorite.”
Three models attempt to project Electoral College results rather than the popular vote. One projects Obama with 213 electoral votes, another with 324 electoral votes (but allows that Romney could win) and a third with 301 votes while acknowledging “a great deal of uncertainty about the outcome.”
Several of these scholars will talk more about forecasting elections on Oct. 16 at the National Press Club. In the meantime, as pollsters continue to track the race, they’ve boldly made their predictions and will await the real results along with the rest of the country.
For previous columns by Dan Balz, go to postpolitics.com.