REX NUTTING
Obama a shoo-in, forecasting models say
Commentary: The economy, the war, and even his face hint at November win
By Rex Nutting, MarketWatch
Last update: 10:31 a.m. EDT July 3, 2008Comments: 296WASHINGTON (MarketWatch) -- Forget about the Rev. Jeremiah Wright or John McCain's war record. If you tell me how the economy is doing now, I'll tell you who will be president next January.
Unfortunately for John McCain's chances, the economy is very weak, and almost everything else is going against him as well, which means Barack Obama will almost certainly be elected president.
That's the view of the overwhelming majority of social scientists who make it their business to peer into the future.
With four months to go until Election Day, the outcome is set in stone, barring some sort of miracle, they say.
The poor state of the economy, the casualties in Iraq, the unpopularity of George W. Bush, the current polling, and Obama's own political skills all point to the election of the Illinois Democrat in November, according to several political scientists, historians and economists who've had a pretty good track record in predicting past elections.
According to their models, it won't be close. Most of them are projecting a 52% to 48% victory for Obama, and that's with assumptions about the economy that are very kind to McCain.
Political scientists noted long ago that presidential elections are fairly predictable because they usually turn on several big issues: How's the economy doing? Is there an unpopular war? Has one party outlived its welcome at 1600 Pennsylvania Avenue? Is one of the candidates a once-in-a-generation leader? Has the current administration done anything outstanding, or scandalous?
It's Obama
There's a whole cottage industry of experts who develop statistical models based on past elections and crunch the numbers to try to predict November's vote now. Almost all of the models say Obama will win.
What actually happens during a campaign -- the ads, the debates, the whistle-stop tours, the endorsements and the innuendo - don't really change the basic landscape that will determine who's in and who's out. In most of the models, it doesn't even matter much who the candidates are, any Democrat and Republican would do as well.
Of course, all these models assume that this year's campaign will be pretty much like those in the past. A major gaffe or stumble by Obama, or superb campaign by McCain could change the dynamics. So could outside events that alter the election landscape.
This election could challenge the models' accuracy, for several reasons.
It's unclear how much Obama's race will influence voters; we've never had an African-American as a major candidate before.
And, for the first time since 1952, no incumbent president or vice president is running. McCain could expose flaws in the models if he's able to distance himself from Bush enough. Most of the models assume that the voters will reward or punish the incumbent party candidate in line with how the incumbent has performed.
What the models say
The granddaddy of the prediction models is American University historian Allan Lichtman's 13 keys to the White House, which include factors such as the economy, foreign policy, scandals, social unrest and even the charisma of the candidates. No one's been elected president since 1860 without holding most of the keys to victory.
By Lichtman's reckoning, seven of the 13 keys are leaning against McCain, and that doesn't even count on Obama being considered charismatic, or on the economy falling into a recession this year. Since polls show the vast majority of voters think the economy is in a recession, we ought to hand Obama that key as well. Perceptions matter more than the technical declaration of a recession.
Yale economist Ray Fair has been at this game a long time too. His model is based on three economic variables, and now predicts McCain will get 48% of the votes. Fair's model had done a good job through the 1988 election, but has drifted further from the actual results in recent contests. You can plug in your own assumptions about the economy and make your own predictions on Fair's website. Go here.
Economist Douglas Hibbs expanded on Fair's idea by including a war variable, which hurts the incumbent party if there are significant casualties in an undeclared war. Hibbs' "Bread and Peace" model explains Eisenhower's victory in 1952 and Nixon's win in 1968. The unpopular war could also be a factor in 2008, but the weak economy is a much bigger reason why the Republicans are likely to get just 48% of the votes this year.
Political scientist Alan Abramowitz of Emory University says his "Time for a Change" model is forecasting a Democratic landslide of about 54% to 46%. But Abramowitz doesn't think it'll be that much of a blowout because there are fewer true independents these days.
"Support for the two major parties has solidified as the parties and their supporters have become increasingly divided along ideological lines," Abramowitz said. "Growing polarization may weaken the effects of short-term forces such as the economy and presidential approval." The approval rating of the president is one of three variables in Abramowitz's model.
Polls, markets and photos
There are other ways of looking ahead to November, including electronic markets, polls, the opinions of experts, and even physical characterisitics, such as height and attractiveness.
Political scientist James Campbell of the University of Buffalo and a few colleagues have created a system that combines several forecasting methods into one. According to the analysis on the Pollyvote.com website updated on Wednesday, McCain is expected to get 47.4% of the two-party vote, the lowest his expected vote has been all year. Read more.
The Pollyvote prediction includes the quantative models of Fair and Hibbs, along with the forecasts of an expert panel, the results of the Iowa Electronic Market, and an average of published opinion polls.
The Iowa market has been shown to outperform the pollsters. In that market, anyone can buy contracts that pay off according to the percentage won by each candidate. As of Wednesday, the market is predicting McCain will get 44.1% of the votes, which is the lowest expected for the Republican candidate in the two-year history of that contract.
Marketing professor J. Scott Armstrong of the Wharton business school doesn't buy the notion that elections necessarily turn on serious issues such as war and peace. In a recent paper, Armstrong and his colleagues found that the perceptions of people who knew nothing about the candidates except their facial appearance did a better job of predicting the outcomes of both the Democratic and Republican primaries than polls published last fall did.
But once again, the news for McCain isn't good: The study found that, based on their faces alone, people think Obama is more competent than McCain.
Rex Nutting is Washington bureau chief of MarketWatch. |
