Are Campaigns Defined by Structure or Agency?

Political scientists will be the first to tell you that campaigns are driven by the fundamentals. The most important variables in an incumbent president’s reelection are if the country is at war and the state of the economy. At the same time, more recent work has found innovative ways that campaigns can induce turnout, cross-pressure voters, and ultimately influence the behavior of voters in a number of ways.

But are these two interpretations at odds? A seemingly innocuous Twitter conversation between journalists and political scientists yesterday casted that doubt, calling it “ironic” in a perjorative sense. The prevailing consensus in that thread: it’s not ironic if different political scientists hold different opinions. But the field at large is still plagued by this impression of cognitive dissonance between campaigns determining elections and fundamentals driving the result. So what is it? Structure or Agency?

@ symbol overload

The easy answer: both.

The better question is: which is more important?

Despite the days I have spent pounding the pavement and knocking on doors, the structure is more important, especially in bigger races. In the sprint to the oval office, fundamentals rule the day. If the US GDP numbers plummet this week, and continue on a downward spin, it won’t matter if Barack Obama could shake the hand of every voter on Election Day and hire the ghost of Ted Sorensen as his speechwriter, he would lose. At the same time, one only has to go back 12 years (or eight, depending on your opinion of John Kerry in Ohio), to find a presidential election that could have been swung with persuasion of a few voters in the right state (Florida).

In political science, the presidential general election is the Super Bowl. The fundamentals are the team’s skill level, experience and general health. The campaign is the coach. So, do coaches win Super Bowls? No they don’t, the players do. Can a coach with ingenious strategy and maneuvers position his players on the field to achieve to their highest potential and win a game they otherwise wouldn’t? Yes. Ask Bill Belichick in 2000.

In statistical terms, these are the reasons we use instruments like a 95% confidence interval, and employ tactics like the margin of error. The best projections and models are still prone to mistakes because ours is a social science, and the behavior of the individual voters is uncertain. Yes, we can project their behavior with a high degree of confidence, but there is a still a measure of uncertainty. No matter how well accurate the structure, there is still a window for agency.

That agency is the ability of campaigns to influence the decisions of individual voters, whether it be to turnout or stay home, or who they choose to vote for. This agency is important. It can determine big elections (see the aforementioned Gore v. Bush, 2000), and it can have an even more important effect as elections get more local. Admittedly, this is painting an important debate in very broad strokes, and I will make sure to revisit this topic in more granular detail in the context of election forecasting, voter learning and the impact of events during the campaigns.