As the primary season comes to a close and the focus turns to the battle between the nominees apparent, it seems like a good time to wonder which, if any, tools have a credible track record in predicting the outcome of presidential races. In his post at Larry J. Sabato’s Crystal Ball, Alan I. Abramowitz argues that, yes, there is one device that has a very impressive track record, and one better than horse race polls — the “electoral barometer.” Abramowitz says of early horse race polls:
The problem…is that they are not very accurate predictors of the actual election results. Polls in the spring of 1988 showed Michael Dukakis with a comfortable lead over George H.W. Bush and polls in June of 1992 showed Bill Clinton running third behind both Bush and H. Ross Perot. So recent polls showing a close race between McCain and Obama may not tell us much about what to expect in November.
There is a better way to go, as Abramowitz explains:
Instead of using early horserace polls, political scientists generally rely on measures of the national political climate to make their forecasts. That is because the national political climate can be measured long before the election and it has been found to exert a powerful influence on the eventual results.
Three indicators of the national political climate have accurately predicted the outcomes of presidential elections since the end of World War II: the incumbent president’s approval rating at mid-year, the growth rate of the economy during the second quarter of the election year, and the length of time the president’s party has held the White House.
…These three factors can be combined to produce an Electoral Barometer score that measures the overall national political climate. The formula for computing this score is simply the president’s net approval rating (approval minus disapproval) in the Gallup Poll plus five times the annual growth rate of real GDP minus 25 if the president’s party has held the White House for two terms or longer. Mathematically, this formula can be written as:
EB = NAR + (5*GDP) – 25.
…A positive Electoral Barometer reading generally predicts victory for the incumbent party while a negative reading generally predicts defeat.
As for the track record, Abramowitz notes:
The Electoral Barometer has predicted the winner of the popular vote in 14 of the 15 presidential elections since World War II. There were five elections in which the Electoral Barometer was negative and the president’s party lost the popular vote in all five of these elections: 1952, 1960, 1976, 1980, and 1992. There were ten elections in which the Electoral Barometer was positive, and the president’s party won the popular vote in nine of these elections: 1948, 1956, 1964, 1972, 1984, 1988, 1996, 2000, and 2004.
Not too shabby. Abramowitz points out that the second quarter GDP figures needed to fill out the electoral barometer formula won’t be available until August. But preliminary stats look very encouraging for Dems:
…Based on President Bush’s net approval rating in the most recent Gallup Poll (-39), the annual growth rate of the economy during the first quarter of 2008 (+0.6 percent), and the fact that the Republican Party has controlled the White House for the past eight years, the current Electoral Barometer reading is a dismal -63.
Abramowitz argues that such an electoral barometer reading would result in a “decisive defeat” for the GOP, which could only be avoided in “an upset of unprecedented magnitude.” The ‘smart money’ crowd at Intrade.com, which also has a more impressive track record than early horse-race polls, seems to agree, with a 61.0 bid for Obama vs. 37.2 for McCain.