Producing a successful presidential campaign requires the coordination of many different pieces, but according to Jon Day, success may be contingent on just a few predictable factors.
Day, an assistant professor at Western Illinois University, gave a talk entitled “The Strategy and Effects of Presidential Campaigns” on Wednesday in the Round Room of the Ford Center for the Fine Arts.
“I thought it was really interesting,” freshman Rachael Morrissey said. “You hear about different campaigns, but you don’t often hear about the actual process of campaigning.”
Day presented a model he has developed that demonstrates the “campaign effect” — the impact that campaigning has on a candidate’s chances of winning the election.
“We like to do Monday morning quarterbacking where we take whoever won the election [and] they were the best campaigner we’ve ever seen,” Day said. “But that’s not always the case.”
Election forecasting models, although usually off by a percentage point or two, are very good at predicting who will win presidential elections. Because forecasts are made before campaigns begin in full force, this suggests that campaigning may have little bearing on a candidate’s chances.
“In most cases, we actually know how the presidential election is going to end up before the election even takes place,” Day said. “If this is the case, this is strong evidence that the campaigns are actually canceling each other out.”
In order to investigate this question, Day represented a campaign as a lottery game between two players. Each player competes for electoral votes and attempts to use limited resources most effectively while his opponent does the same.
Studies show that campaign activity follows a trend of diminishing returns: each additional ad, for instance, reaches and converts fewer voters. The question therefore becomes at what point candidates should stop spending, all the while knowing that the other side is facing the same concerns.
According to Day’s model, campaigns will reach an equilibrium point in terms of spending, targeting resources in the same states and rendering their efforts moot.
Day illustrated the importance of keeping up with the spending of one’s opponent via the example of Florida in 2000, where Al Gore lost by 537 votes and consequently lost the entire election. Day attributed Gore’s loss to his having underspent George W. Bush by $10 million in advertising.
“If he would have spent in equilibrium, he should have received 41,742 more votes, according to my model,” Day said. “He would have won the election.”
Another factor at play is how good a candidate is at persuading voters. This “persuasiveness” aspect determines how campaigns actually spend resources. Because more persuasive candidates are better able to convert voters, they should spend more resources. Even though they will experience diminishing returns, each stop on the campaign trail will be more effective than stops made by the less persuasive candidate.
The problem, then, comes with measuring persuasiveness.
“Political scientists are still working on how to do that,” Day said.
At the conclusion of his talk, Day pointed out that his model is not dynamic. In the future, he hopes to expand his model so that political scientists will be able to see how gaffes influence campaign effectiveness on a week-to-week basis.