If you ask experts, Donald Trump's plan to ban Muslims from entering the United States would be virtually impossible to implement. The wall on the Mexican border? Same. Mass deportations? Same.
Trump -- like many politicians -- has a tendency to make grandiose, unachievable statements meant to bolster his electoral chances. Trump's just happen to be more grandiose and less achievable than many, because he has learned that there's little downside to doing so. Of course Mexico isn't going to pay for a wall on the border, but Trump keeps saying it, and people keep liking him for saying it will happen.
It's worth keeping that in mind when we consider Trump's recent argument that he worries the results of the general election will be rigged. After Mitt Romney lost in 2012, Trump complained on Twitter that a "revolution" was needed because the election was "a total sham and a travesty." (Since he began running for president, the blame for Romney's loss has been shifted onto Romney, whom Trump dislikes and vice versa.) The Post's Dave Weigel explored the sketchy, disconcerting politics of Trump's argument on Wednesday, noting that it plays into Trump's me-against-the-system rhetoric. But it's worth establishing another aspect of it: It's simply not feasible.
Let's work backward. In order to win the presidency, you need to win a majority of the electors who vote in the electoral college. There are 538 in total, necessitating the support of 270 electors. How those electors vote is tied to the results of the election in a given state. Win California, you win the voters of California's electors. This isn't uniform. Maine and Nebraska divvy up their electoral votes based on the results in their congressional districts. But that's mostly how it works.
On Wednesday, one of the members of the electoral college from Georgia made an unusual revelation, saying that he wouldn't support Trump if the businessman won the state. He's technically allowed to do this, under the state's rules; 21 states don't formally bind electors to the results. As a result, there have been a number of so-called "faithless electors" in the past, who have cast votes they weren't supposed to or who have abstained, though none have changed the results of the election.
Trump uses the word "rigged" in a variety of ways. In his interview with The Post's Philip Rucker on Tuesday, he lumped several different things together, including his annoyance at unbound delegates in Louisiana during the primaries and vague allegations of voter fraud. If Hillary Clinton conspired with the electors in those 21 states to get them to back her no matter what, even if she lost, you could certainly say that the electoral vote was rigged or stolen -- but it wouldn't be a secret. It is hard to imagine a scenario in which the result wouldn't be somehow contested.
What Trump is clearly implying is that the election might be secretly stolen. That's why he brought up vote fraud in that interview, saying that "there’s a lot of dirty pool played at the election, meaning the election is rigged." (His longtime ally Roger Stone discussed the idea rather ominously in an interview last week, saying that results not in line with polling would result in "a constitutional crisis, widespread civil disobedience, and the government will no longer be the government.") So what would secretly stealing an election take?
We saw in Florida in 2000 that a close result in a key state can tip the result. A thousand votes in that state for Al Gore and he's the president sworn in on Jan. 20, 2001. This is a go-to example for conspiracists, understandably.
But it's important to remember that you can't predict which state will be key. If you're going to rig the vote, you need to do it in a number of places at once -- which increases the risk, complexity and number of people involved. Adding a thousand votes in Florida would have made the difference, but that's only because George W. Bush won enough votes in other states to get him close to 270. You need to be able to predict the results in every swing state, or you need to rig votes across a broad geography. That's far harder than it seems at first blush.
This bring us to the process of stealing votes. "We may have people vote 10 times," Trump said to Rucker, by way of explaining the risk of fraud. Trump's voting record is patchy, but he's probably seen the restrictions that apply to voting in New York, at least. You have to be registered, and you have to sign a book, which election workers compare to the signature on file. In many places, you can vote provisionally if your records aren't found, but if you vote multiple times in multiple places, those provisional ballots will not be counted. On CNN on Tuesday, former Trump campaign manager Corey Lewandowski defended Trump (as he does) by saying that in his home state of New Hampshire, anyone can simply go vote, if they live there or not. It's not that easy, of course. You have to be able to prove your age and place of residence.
That it's cumbersome is the point. That's a big part of why an investigation of 14 years of elections found only 31 cases of in-person voter fraud. That's a minute fraction of the 1 billion ballots cast.
We made an interactive showing how rare in-person fraud is. Click the button below. Every 29 hours that it's running, it should find one example of fraud.
But we keep losing sight of scale. Bush won Florida by about 500 votes. Under Trump's 10-votes-per-person formulation, that's 50 people who would need to go and vote fraudulently 10 times and get away with it. Fifty people -- if you know the results otherwise and everywhere else. If you are loading the dice in every swing state, you need a lot more people than that. Then you need people to coordinate those people. We're talking about a cast of hundreds, all of whom would have to get away with the conspiracy undetected. Building a wall on the Mexican border sounds easy: Building a wall isn't hard! But when you consider scale, you realize how difficult it becomes.
A number of people who haven't read this far have already emailed me to talk about voting machines. So let's talk voting machines.
Two things made the idea of Bush stealing his reelection in 2004 palpable to anxious liberals. The first was the memory of the conclusion to 2000, which is fair. The second was the fact that executives from Diebold, which made voting machines, were supporting Bush in the election. Why not simply crack open the machines or do a hack or whatever and change the results after people voted? There were stories (that have been repeated since) of people pushing the button for one candidate and then seeing a vote cast for the candidate's opponent. Conspiracy.
But, again: scale. Hacking voting machines (which is certainly possible!) would make it easy to add a lot more than 10 votes in particular precincts. Those machines aren't distributed from some centralized Evil, Inc. warehouse before Election Day, though. They're maintained by counties across the United States. Here's a list of the machines used in each county in New York, for example. Here's how to vote in every county in California. Here's a look from last year at the diverse system of machines in use across the country. Hacking this ad hoc system to skew the results requires hacking a lot of machines with different code in a lot of places without detection -- and doing it in person, since they're not network-connected. That again means a lot of effort and a lot of risk.
There's one other point at which the results can be tweaked. Once votes are cast, they're tabulated, usually at a county level, and passed on to the state. You can see where this is going. Sure, someone working for the board of elections in Manhattan could throw an extra 1,000 votes on Hillary Clinton's tally, but that person would have to do so without it being noticed and without confidence it would make a difference. To ensure victory, you'd need lots of people in any number of counties across the country who could get away with it. Even if it's just two dozen counties in key swing states, that's a lot of government employees to persuade to violate the law.
On top of all of that, the pattern of vote-rigging would itself have to escape notice. Vote results often follow patterns that people watch on election night to try to predict the outcome or after the fact to assess why an election happened the way it did. To keep such a thing from being noticed, you'd need to be very careful about where and how much you inflated the tally. One pattern so far is that Hillary Clinton is leading in national and state polls consistently, which is another reason that her rigging the election seems unlikely.
It's probably clear from this that stealing a local, smaller-scale election would be far easier. If scale's the problem, reduce the scale. No sweat. Carrying this off at a state or national level is a different ballgame.
There are ways to shift the vote results in a state, of course. Last week, one such proposal was uncovered and rejected by a federal court. In that plot, black voters in North Carolina were targeted to try to suppress the number of votes they cast. Who was behind the scheme? Legislators who passed the sort of voter ID law that Trump, in his interview with The Post, claimed was necessary to keep people from voting 10 times. A study of turnout from 2012 found that voter ID laws in Kansas and Tennessee dropped the number of votes in those states by 122,000 -- heavily among black and younger voters, who typically vote Democratic. Why hack a machine when you can hack the process?
It's possible that Trump sincerely thinks that the results of the November election are at risk from some coordinated effort to throw the vote. And it's theoretically possible that it could happen, just as it is theoretically possible that Mexico would pay billions of dollars to build a wall on American soil. But it's pretty obvious that neither will happen -- and that pretending each might is valuable to Donald Trump.
To rig an election, you would need (1) technological capabilities that exist only in Mission Impossible movies, plus (2) the cooperation of the Republicans and Democrats who are serving as the polling place’s election officials, plus (3) the blind eyes of the partisan pollwatchers who are standing over their shoulders, plus (4) the cooperation of another set of Republicans and Democrats — the officials at the post-elections canvass, plus (5) the blind eyes of the canvass watchers, too. Then you’d still have to jedi-mind trick lawyers, political operatives and state election administrators, all of whom scrub precinct-level returns for aberrant election results, and scrutinize any polling place result that is not in line with what they would have expected, based on current political dynamics and historical election results.