The Importance of Unemployment Trends
by David Atkins
Take this with a grain of salt (or several handfuls), but the implications of the model are certainly interesting:
A couple of economic researchers working for Yahoo say they have developed a forecasting model that predicts a win for President Barack Obama this November.
The model, created by David Rothschild and Patrick Hummel, predicts that Mr. Obama will carry 303 electoral votes this fall. At the same time, they say that several states contests remain virtual toss-ups.
In reviewing the last ten presidential cycles, the researchers say that their model correctly predicted the eventual winners in 88% of the 500 state elections that took place. Between now and mid-June, they they assume that personal income growth remains average for a reelection year and that the president's approval rating remains at or above its current 48% range.
One aspect of the model that might generate controversy within the ranks of professional politicians and their handlers is the conclusion about the value of a campaign and a candidate: Rothschild and Hummel believe that it matters far less than conventional opinion has conventionally believed.
"One of the interesting findings of the research is, quite frankly, that you can predict outcomes of elections with pretty amazing accuracy pretty far away," Rothschild said.
The researchers found that an increase or decrease in unemployment trend lines was a much more powerful predictor of election results than the unemployment rate itself. In particular, a key data point was the state-by-state growth in income in the first quarter of the election year. With the main economic indicators trending up, albeit at historically depressed levels, they say that's an encouraging harbinger for an Obama victory.
"The net effect of campaigns are meaningful but not massive," Rothschild said. He said that the economy's first and second quarters (in an election year) "more strongly correlate with a president's reelection chances."
I dislike and distrust the notion increasingly trendy in academic and church-of-the-savvy sources that campaign structure and rhetoric make little difference to electoral outcomes, which are supposedly predetermined by a combination of voter registration, name ID, approval polling and economic numbers. That's certainly not true of more local elections, though it may be more reflective of presidential campaigns. At the very least, campaigns and rhetoric influence the political ground on which the post-electoral legislative battles are fought--which is the only good reason to care about politics. Without the real implications for consequences in people's lives, the horse race of political elections is a dreary business, a reality show far more corrupt and less entertaining than most sporting events.
Still, insofar as the model has credibility, it shows the importance of maintaining positive employment trends throughout the first half of a presidential election year. The Obama Administration would do well to keep that clearly in mind.
.