What happens if the election is close?

By Clark Miller

As I write, Nate Silver has updated his 538 model to show a 92.2% probability that Obama wins easily. The reason? His model calculation concludes that Obama will win Ohio with a 3.8% margin, Nevada by 4.7%, and Wisconsin by 5.7%. With those three states, Obama’s Electoral College total will be 271. Assuming the polls are right, this election will not be close. Moreover, if Obama’s victory is by these margins, we’ll know the winner by 7 pm ET.

My colleague, Gregg Zachary, is convinced that, having learned the lesson of Al Gore, neither side will concede this election. I see his point, but politics is still a game of credibility. The only scenario in which either side will be able to credibly avoid a concession speech is if the election outcome is potentially up for grabs in one or more decisive states. In realistic terms, this would require an election margin less than 0.5% in one or more states whose votes are necessary to put either candidate over 269 Electoral College votes. That has happened only very rarely in US history, and it’s unlikely to occur today.

But what if it does? A decade ago, I spent a fair bit of time deconstructing the 2000 presidential election dispute and wrote an article about it that you can find here. Here are some lessons I learned along the way:

  • This may seem obvious (but somehow it always gets left out of introductory textbooks): democracy depends utterly on the ability to construct facts. The electoral system is a complex, highly networked socio-technical knowledge system for producing those facts. Its insides are not pretty, but they are designed to accrete credibility around one particular fact—who won the election—slowly, over time, but with confidence.
  • This is crucial: an exact, objective count is unnecessary. All that matters is an objective determination of who won. The courts have been clear forever that they will not open up an election for scrutiny unless enough votes are in contention to plausibly alter the outcome of the election. The courts are also clear that they will not cast out ballots that have been cast substantially correctly over minor rule violations. The right to vote is too important.
  • This is equally crucial: Zachary has put his finger on a sensitive point. Historically, the ritual of concession speeches has been important in conferring facticity. Big tobacco refused to concede the health risks of smoking for decades. Republicans have refused to concede the fact of climate change for almost as long. Absent a concession, determination of fact is much, much, much harder.
  • This is also crucial: at the end of the day, government is a fiction of our collective imagination as American citizens. You don’t even have to generate an objective winner of the election in order to have a de facto legitimate government, if the American public chooses not to contest the outcome. 2000 proved this. In other words, the American people’s collective sentiments form a backstop for the electoral system that can sanction (or not), should they so choose, a President.
  • Final crucial point: There is not actually a presidential election. There are, rather, fifty separate state elections, according to fifty different state laws. Zachary doesn’t like this, but in point of fact, it makes the system more stable, not less. It’s counterintuitive, but imagine the opposite case: a national election, with the winner based on who won the popular vote. Now imagine a race separated by 0.1%, well within the margins for an automatic recount. We’d have to recount every ballot, in every state, with no one able to monitor every nook and cranny of the system. And what if it went to a contested court case? Every irregularity in every polling place would potentially be relevant and so would have to be adjudicated. Instead, in 2000, we only had to fight out a single state in the courts; much easier and much more contained. It’s a trade-off, but one I’m comfortable with. (Zachary thinks that if we just had a national election system like Germany, everything would be fine—and he’s not alone in this view. But the real reason their system works without rancor is because in Parliamentary elections, a few thousand votes either way really doesn’t matter. So what if the margin is 242-200 instead of 243-199 seats in Parliament. Even 222-220 versus 221-221 isn’t that big a deal, although it may determine who gets the first shot at forming a government. In our system, however, a few votes can make the difference between four years of ultra-liberal versus ultra-conservative policies from the most powerful institution in society.)
  • The process of an electoral contest generally goes something like this: (1) an election count will be determined; (2) if the difference falls within a certain range, usually around 0.5%, an automatic recount will be generated; under other circumstances, candidates may request a recount; (3) once a recount has been finished, an official election count will be certified; (4) once the count has been certified, a candidate may contest the election in court, if they have observed irregularities that call into question the official count; (5) the courts will hear the evidence in the cases and rule; (6) those rulings will be subject to appeal, either through the state court system or through the federal court system, ultimately ending either in the state supreme court, the US Supreme Court, or both. Keep in mind that there are five key parties in the system: the two campaigns, the state election administration, the state court system, and the federal court system.
  • There is a timetable. Unlike other elections in the US, there is a clear timetable for resolution of a disputed presidential election. Congress will meet in joint session on January 6, 2013 to count the votes of the Electoral College. Prior to this, on December 17, 2012, electors in each state meet to cast their votes. In order to be official, states must appoint electors by December 11, 2012. Hence, states must resolve any controversies in electoral tallies prior to this date, or else they risk losing their votes in the Electoral College.
  • If an electoral contest occurs, it will be an unholy mess (if, thankfully, isolated to a relatively few states; see above). If one occurs, we will learn far more about the electoral laws of Ohio than we’ve ever cared to. Everything will be potentially contestable. Indeed, the legal contest is already under way. Already, democrats have sued both to prevent the Ohio Secretary of State from inappropriately reducing the times allowed for early voting and to question the addition of software (allegedly that has been inappropriately reviewed and tested) to Ohio voting machines. We will get to learn whether the great MIT-Caltech solution to voting machine problems from 2000—voter verified paper audit trails—actually works to reduce controversy over electronic vote counts (my bet, if the election is contested: they won’t).
  • In the face of this mess, have patience. Both election administrations and the courts have every reason to pursue credible outcomes. Moreover, they are all backstopped, meaning all decisions will be appealed to and looked over by another node in the network. Only the Supreme Court has no one to get their back, which makes me wonder whether the Roberts Court, which has often seemed to want to avoid highly partisan appearances, would actually wade into a dispute this year as forthrightly as their predecessors did.
  • The media, as a socio-technological lens that the public will use to view the controversy, will play a critical role in shaping the credibility of electoral counts. How the media will play a controversy is unclear? Given 2000, as well as the ongoing electoral disputes in subsequent years, will media experience lead to less sensational and confused coverage than 12 years ago? Or will the media interest in a story override learning and lead the media to exacerbate the controversy, at least in its first few days and weeks?
  • The single most important element in maintaining long-term confidence in the US electoral system is maintaining public trust in the human elements of the election system. The people who work in election administration are crucial to creating credible counts of votes. Unfortunately, these people took a huge beating in 2000, and subsequent events have only continued to cast doubt on their ability to carry out free and fair elections. This year, already, several secretaries of state (including in Ohio and Florida) have made highly partisan decisions designed to impact voting outcomes. This makes them appear biased and calls into question the trustworthiness of their institutions. Any election controversies will further undermine public confidence.

Clark A. Miller is associate director of the Consortium for Science, Policy & Outcomes and chair of the program in Human and Social Dimensions of Science and Technology at Arizona State University. An engineer by training, he writes on science and democracy and the politics of energy transformation.