Election Forecast Models Clouded by Economy’s Slow Growth
Moody’s Analytics says its economic- forecasting model shows President Barack Obama winning re- election with 303 electoral votes. Economist Douglas Hibbs Jr., pointing to slow income growth, predicts a Mitt Romney victory. To Yale University’s Ray Fair, the race is simply “too close to call.”
While the state of the U.S. economy is the No. 1 issue in the election, even nonpartisan forecasters who crunch data and claim to have predicted the results in past presidential races can’t agree on what impact it will have on the outcome.
Sophisticated models in the past have bolstered arguments that presidential elections are mostly predetermined by the economy. This time, some concede that the plodding recovery may not tilt the balance toward either candidate. Instead, the costly, cross-country political campaign, which these forecasters have dismissed as irrelevant, may tip the balance.
“When you get down to that point, where I’m saying that it’s going to be decided by less than one percentage point, then it’s possible that the campaigns really do matter,” said Alan Abramowitz, a political science professor at Emory University in Atlanta who developed one of the most-followed forecasting models.
U.S. companies added more workers than forecast in August, a private payroll services company reported this morning a day ahead of an official government report, offering hope that the labor market is improving. Roseland, New Jersey-based ADP Employer Services reported a 201,000 rise in jobs, exceeding the 140,000 median estimate of economists surveyed by Bloomberg.
The economy’s condition heading into the election is clear right now: a slow recovery from the worst U.S. downturn since the Great Depression.
Economic growth this year is below normal for the post- World War II era, though stronger than when Dwight Eisenhower was re-elected in a 1956 landslide. Even with unemployment at its highest in any election year since the war, joblessness has declined more over the past 12 months than during the same period of any postwar election year except 1984 and 1976.
What matters most to voters isn’t the absolute state of the economy, it’s the trajectory, said Christopher Wlezien, a professor at Temple University in Philadelphia and co-author of the forthcoming book “The Timeline of Presidential Elections.”
The election models point in different directions based on what their developers say influences voters the most.
The Moody’s electoral vote projection, which assesses economic performance state by state, gives Obama the advantage, reflecting better circumstances in many of this year’s battlegrounds.
Obama wins Ohio, where the unemployment rate was 7.2 percent in July, below the 8.3 percent national average that month, and Virginia, where joblessness was at 5.9 percent, under the Moody’s projection. Romney picks up Florida, where unemployment was 8.8 percent in July.
The Republican challenger can expect victory under the “Bread and Peace” model developed by Hibbs, a retired professor who taught at Sweden’s University of Gothenburg. He says the drivers of voter opinion are war casualties and income growth. He forecasts Obama will receive 47.5 percent of the two- party vote to Romney’s 52.5 percent. The Afghanistan war has produced few casualties relative to the Vietnam and Korean wars.
Yale’s Fair, who developed one of the first models to forecast elections based solely on economic indicators, projects Obama will win 49.5 percent of the two-party vote to Romney’s 50.5 percent, based on gross domestic product growth and inflation under the incumbent. That amounts to “too close to call” given an error margin of three percentage points, he said.
Two models that weigh public opinion polls as well as economic data give Obama the advantage.
Abramowitz’s “Time for Change,” based on GDP growth during the April-to-June quarter and presidential job approval in the Gallup Poll for the last three days of June, anticipates an Obama win with 50.5 percent of the two-party vote, with a 1.5 percentage-point margin of error.
While some of the election modelers say polling data detracts from the goal of capturing how fundamental forces shape the vote, Abramowitz said he includes public opinion because “you get a much better forecast.”
Wlezien and Robert Erikson, a political science professor at Columbia University, look to leading economic indicators through the first three months of the election year and an average of all polls of voter preferences. Based on polling data for July, they predict a bigger Obama win, with 52.6 percent of the two-party vote and a 3.1 percentage-point margin of error.
Poll results move closer to the outcome suggested by the economy’s condition as the election year progresses and “voters increasingly take stock of things,” Wlezien said.
The forecasters have an imperfect record. Most of the major forecast models predicted the election of Democrat Al Gore in 2000. In most cases, developers say, the models are designed to predict the popular vote, which Gore won with 48.4 percent to George W. Bush’s 47.9 percent.
Designing models to forecast elections is challenging because of the limited record, Phillips said. There have been only 16 presidential elections since World War II. Seven of 10 incumbents have been returned to office. While economic growth on average has been higher than normal during the years in which incumbents won re-election, strong growth hasn’t necessarily translated to re-election nor weak growth to defeat.
GDP, which has grown at an average 3.2 percent annual rate since 1947, expanded 0.3 percent during the first nine months of 1956, when Eisenhower was re-elected with 57 percent of the vote. George W. Bush won in 2004, following average 2.8 percent growth in GDP during the first three quarters of the year. It has averaged 1.85 percent during the first half of 2012.
Skeptics of election models that rely on economic indicators say the issues that excite voters are constantly in transition, be it war, terrorism, cultural division or fear of crime.
Some models make projections based on more than just the economy. An Obama win is forecast by the “13 Keys” developed by Allan Lichtman, a history professor at American University in Washington, and journalist Ken DeCell. The keys include economic growth and control of Congress, as well as domestic and foreign- policy achievements and the lack of scandal or social unrest. Nine of the 13 keys have turned in favor of re-election.
Bloomberg reserves the right to edit or remove comments but is under no obligation to do so, or to explain individual moderation decisions.