The question for the future is whether, in close presidential elections, there will be a significant chance of the lead changing hands in several states, as in 1960—or instead whether the leads will stay with the same candidate throughout the canvass despite the inevitable adjustments in the vote totals during that period, as occurred in 1976 and 2004, when the presidential elections were also close. The answer depends on just how large the big blue shift in the canvass has become, and on the extent to which the extremely large gains that Obama made during the canvass in battleground states is a function of factors other than his overall national popularity. As I have indicated, further insight on just what is causing the observable big blue shift must await more sophisticated statistical analysis.

Meanwhile, it is worth considering whether it would be possible for the media and the public to become comfortable with a new normal in presidential elections where there is a significant chance of leads in states changing from one candidate to another after Election Night and during the canvass. Perhaps it will be no big deal if, for example, on the morning after Election Night the Democrat is behind in California and Washington, by several hundred thousand votes in each, but everyone expects the Democrat easily to overcome this deficit during the canvass through the counting of as-yet-unprocessed absentee ballots. Indeed, the media might be prepared to call these states for the Democrat even as the official (albeit incomplete and uncertified) returns have the Republican in the lead. Moreover, in this situation there might be no threat of litigation because everyone would know that the anticipated flip in the lead is just a normal part of the vote-counting process with so many absentee ballots still uncounted. Even though the presidential election would technically remain unsettled in overtime, in a sense this situation is no different from any other: The election is always officially unsettled prior to certification, and if both candidates treat it as a foregone conclusion that the Democrat has won these states, then the race there will feel over just as much as any other called by the networks on Election Night.

But I surmise that solidly blue states like California and Washington are very different in this respect from battleground states like Ohio, Florida, and Colorado. If the Republican is ahead in these states by tens of thousands of votes on the morning after Election Day, the networks will not be so quick to call the states for the Democrat, even though it is anticipated that the Democrat is likely capable of making up most, if not all, of those deficits. Moreover, the Republican candidate will not quickly concede those states to the Democrat, and if these states are essential to winning an Electoral College majority, the threat of litigation will be very much in the air. Indeed, by the morning after Election Night, the litigation over as-yet-uncounted ballots in these crucial states might already be underway. When the election going into overtime is unsettled in battleground states, unlike in solidly blue or red states, overtime means litigation rather than waiting for the completion of the canvass in a harmonious manner.

Therefore, battleground states need to prepare themselves for the possibility that Election Night might come and go with the candidates realizing that it is very much worth fighting for votes during the canvass. The magnitude of this risk is the question addressed through the examination of the data presented here. At this point, it remains unclear whether the margin of litigation for battleground states is closer to a 25,000-vote deficit for the Democrat, which is approximately the level of the 2004 election, or instead closer to 75,000-vote deficit, which is more like what we saw in 2008 and 2012. Further statistical analysis of the data may help refine the assessment of this risk. But in the absence of this additional information, it would be prudent for battleground states to consider changes to their voting rules and procedures that might help to reduce this risk. These states would also do well to specifically scrutinize their rules and procedures for the canvassing and recounting of ballots—their “overtime” processes, so to speak—to evaluate the extent to which they are ready to handle the intense pressure that will be put on these overtime processes if indeed one of their elections falls within what may well be a rapidly increasing margin of litigation.