Who Will Be Inaugurated on January 20? (Almost Certainly Donald Trump)

Recent days have seen a surge in efforts to change the 2016 presidential results. Jill Stein is spearheading calls for recounts. Democratic-affiliated electors are joining attempts to convince their Republican-affiliated counterparts to vote faithlessly on December 19. In response to questions about these post-election developments, I thought I would provide a few quick thoughts. The punchline is straightforward: none of this activity is likely to change who will become the 45th president. The following discussion provides more context.

First, for those trying to understand the recounts, state law is what governs. As a result, the rules governing recounts vary, and the answers to legal questions—including when and how to demand a recount and which standards govern the proceedings themselves—depend on whether one is seeking to recount votes in Wisconsin, or Michigan, or Pennsylvania. Despite such variation in the rules, the ultimate outcomes of these recounts are likely to be the same. Because taken either together or individually, they are very unlikely to make a difference in the 2016 presidential race.

By my count, the outcome of the 2016 presidential race changes only if recounts flip the results in all three of these states. This is because Donald Trump’s current elector total is 306, which means he has 36 more than he needs. In other words, Trump would need to lose the votes of 37 electors in order to drop below 270. Wisconsin provides 10, Michigan provides 16, and Pennsylvania provides 20. (There is, of course, the possibility that some of the Trump electors—that is, some of the electors who were selected based on Trump’s statewide victories—will prove to be “faithless,” thereby dropping Trump’s totals under 270 without all three of these states necessarily being flipped. See below for why this is unlikely to occur and why, even if it did occur, it is unlikely to change the result.)

What are the odds that recounts will change the outcomes in all three of these states? Based on historical evidence, the odds are minuscule.

In recount terms, the apparent margins of victory in the 2016 election are huge. As a result, there is not a single modern precedent for even one of these states to experience a reversal. Supporting the conclusion that there is “essentially zero chance” that recounts in these states will reverse Trump’s lead, Josh Douglas observes that, in the last 15 years, a statewide recount has flipped the outcome of an election only three times. Each of these reversals occurred in a really close race. According to FairVote, the first of these unicorns was spotted in 2004, where a recount in Washington State swung the margin by 390 votes, which translated into 0.014% of the votes cast. The second emerged in Vermont in 2006, where the recount swung the margin by 239, or 0.107% of the votes cast. The third could be found in Minnesota in 2008, where the recount swung the margin by 440 votes, or 0.018% of the votes cast. As FairVote concluded, based on its comprehensive analysis of all the recounts (consequential or otherwise) over a ten-year period, recounts tend to “change the margin by insignificant numbers.”

The problem for Stein and others hoping for game-changing recounts? The 2016 presidential vote totals do not turn on insignificant numbers. Current estimates indicate that in Wisconsin, Donald Trump is ahead by over 20,000 votes (equaling approximately 0.8% of the total votes cast); that in Michigan he is ahead by over 10,000 votes (approximately 0.3% of the total); and in Pennsylvania he is ahead by over 65,000 votes (approximately 1.2% of the total). The idea that standard-issue recount proceedings would flip the outcome not only in one of these states, but in all three, seems beyond the realm of possibility. This reality might help to explain why Marc Elias, the lead attorney for the Clinton campaign, has repeatedly insisted that “Hillary Clinton’s campaign didn’t want this recount and doesn’t think it will change anything.”

Some nevertheless have suggested that illegal hacking—rather than innocuous errors—might explain Donald Trump’s leads. Under this theory, the prior recount precedents are not on point. The trouble with this theory is that there appears to be no compelling evidence to back up the hacking claims—and under any of the states’ election laws, unsubstantiated theories about the possibility of hacking (or other forms of fraud) are far from enough to overturn the results. This is, incidentally, a very good thing; the democratic process is threatened by baseless accusations of election malfeasance.

In short, it seems close to certain that recounts in Wisconsin, Michigan, and/or Pennsylvania will fail to change the outcome of the 2016 presidential race. In making this assertion, it’s appropriate to acknowledge that many observers of the 2016 elections placed far too much faith in historical precedent and various forms of statistical analysis when predicting how the election itself would turn out. And it’s important not to do the same here. Still, given the wide margins in these three states (and the fact that a “recount” is, at core, simply a reconsideration and re-tallying of an already analyzed set of ballots), it really is hard to understand how the ultimate outcome of the presidential race possibly could be flipped.

This, finally, leads to the question of whether a separate effort—that is, the effort to convince sufficient electors, among those who were selected based on Trump’s statewide victories, to go rogue and vote for someone other than their candidate—has any chance of changing the result. Again, the magic number is 37; anything less than that, and Trump still has the 270 votes he needs. (And, to be clear, the 37 electors need to come from states that went for Trump; it doesn’t change anything if, for example, a Washington State elector carries through with his loudly proclaimed promise not to vote for Hillary Clinton.) Among the 306 Republican-affiliated electors who have signed up to vote for the Republican candidate, is it possible that over 10 percent of them will refuse, in the face of clear election results, to cast a vote in favor of their state's preferred candidate?

Such a development is not literally impossible, and it is true that at least one such elector (from Texas) already is refusing to vote for Trump. But this outcome again seems exceedingly unlikely. Remember that these electors are not random voters or dispassionate observers; quite to the contrary, these are people selected through Republican party apparatuses to be entrusted to vote on December 19 for the candidate selected by the party. Moreover, refusing to vote for Trump would not be the same as deciding, in a vacuum, which candidate happens to be the elector's preferred choice. Rather, it would require each of these political operatives to disregard the will of their own states’ voters, who just a few weeks ago voted not for Hillary Clinton, and not for some third party candidate, but for Donald Trump. To this end, it is telling that the Trump-averse elector from Texas will not be voting for another candidate on December 19; rather, he has resigned his position as elector, thereby allowing someone who is willing to vote for Donald Trump to replace him.

And here’s the kicker to all this: even if dozens of Republican-affiliated electors were indeed to refuse to vote for Trump, thereby reducing his total to under 270 electoral votes, even that likely would not stop him from taking office. Rather, the 12th Amendment requires that, in the absence of any candidate receiving 270 votes, the decision then go to the House of Representatives. Which will be Republican controlled. And which therefore, in all likelihood, would then vote for (you guessed it): Donald Trump.

What's more, as Ned Foley has pointed out, a bizarre and convoluted statute—the Electoral Count Act of 1887—very well may empower Congress to reject rogue electors’ votes even without the 12th Amendment backstop. Without getting too much into the (deep, disorienting) weeds, the Electoral Count Act seems to empower Congress to choose between competing claims over electors. In the face of Trump challenging votes cast by faithless electors, and with Republicans set to control both the Senate and the House, it seems highly likely that Congress would avail itself of this option.

Assuming (notwithstanding all indications to the contrary) that Congress would not take such a step, it is true that there does exist a route for avoiding a Trump presidency. In this scenario, at least 38 Republican-affiliated electors (note that an additional elector, in this scenario, would be necessary) would need to cast their votes for Hillary Clinton, rather than for Donald Trump, or for some other candidate, or for no one. At that point—again, assuming Congress simply were to accept such an unprecedented and controversial result rather than to resist it via the Electoral Count Act—that would bring Clinton’s electoral vote total to 270, and she would be entitled to the Presidency. A similar outcome would adhere if at least 270 (!) electors collectively broke ranks and voted for what faithless-elector advocates are referring to as a "compromise candidate." It is an understatement to say that neither of these developments seems practically or politically realistic, particularly when their success would require the tacit acquiescence of a Republican-controlled Congress.

I have one final point concerning the idea that sufficient electors might break ranks to change the outcome. In the exceedingly unlikely case that the scenario somehow were to come to pass, it is hard for me to imagine what the reaction might be from the approximately 62 million people who voted for Trump—that is, from the approximately 62 million people who voted for the candidate who seemed to have won the election but, as a result of faithlessness on the part of a handful of political insiders, suddenly lost the presidency. For those who suggest that this development could garner legitimacy—for reasons that purportedly are candidate-neutral—I would invite them to consider what their own reactions might be if the shoe were on the other foot. If an Election Night victory for Hillary Clinton, when set against the faithlessness of a relatively small number of electors, turned into a Donald Trump presidency, would they accept such a result? Do they really predict that Trump voters would?

All of this is not to say that these post-election efforts have no value. To the contrary, they constitute a protest of sorts by those who are deeply concerned about the results of the 2016 presidential election. Still, for various practical, political, and legal reasons, these efforts remain exceedingly unlikely to change which candidate, come late January, will be facing the Chief Justice with one hand in the air.