This is a weird piece for me to be writing. I’ve been following the field of voting for a little while (close association to people working for ACCURATE sparked my interest), so when I see these articles about a rigged election via vote stuffing and voters getting bused around from precinct to precinct, I roll my eyes out of habit. There are just so many better ways to steal an election! Hacking the machines can take mere minutes, fewer if you can use a USB stick, corrupt election officials can control the process from start to finish…
Still, there seems to be real worry about it out there, so I figure it’s important to explain exactly why this doesn’t happen in large numbers. In-person voter fraud does happen, of course; Melowese Richardson was convicted for it in 2013, and Terri Lynn Rote was charged just a few days ago. The thing is, it’s just not practical or cost efficient. I hope to prove this to you by way of a thought experiment.
Note: all the numbers used in here were found using simple Google searches and reading a bunch. I haven’t linked to sources since they’re easy to find in obvious places (e.g., the North Carolina State Board of Elections website).
Ground rules
If we’re going to look at this thing, we need to set some parameters. First off, we’re going to give this voter fraud scheme the best chance we can, meaning that we’ll be rounding in whatever way is most favorable to it succeeding. Next, we’re going to assume that we’re looking at a presidential candidate who wants to win by flipping at least one state via in-person voter fraud, and that this person wants to get away with it. We’re also going to assume that there’s an organized effort to get people to do this; just hoping that tens of thousands (yes, it’d take that many) of people in the right place will break the law for one candidate’s benefit is not something a campaign can count on.
We’ll assume a few things about our voters too. We’ll assume that they’re fired up party loyalists who don’t require payment, and are fired up enough that each one has a 100% chance of voting. This is obviously outside the realm of plausibility, but as I said, we’re giving this thing the best chance we can, and besides, accounting for shrinkage is more math that we don’t need to do just to prove this is a bad idea.
Finally, we’re going to focus on a single state. The stuff we’ll be looking at transfers to just about any other battleground state, with the same challenges to overcome.
Picking a state
This is the most crucial step of the process. There’s absolutely no reason for Trump to try flipping California, or for Clinton to try flipping Oklahoma. You’d just need too many people — and “too many people” is going to be a recurrent theme in this piece. For my money, North Carolina is the best state to target right now. Sure, Florida and Ohio have more Electoral College votes, but FiveThirtyEight lists North Carolina at #4 in both tipping point and voter power index rankings, meaning that it’s a good bet that extra votes here will be extra powerful.
North Carolina has about 6.6 million registered voters, about 70% of whom show up to vote. This comes out to about 4.5 million votes in the state. Unfortunately, higher turnout means you need more in-person fraudulent voters to tip the election, so to help the scheme along, we’ll round down to 4 million votes even.
How many fraudulent voters?
Now we need to decide how many of our fired-up party loyalists to mobilize. Ideally, we’d like to mobilize as few people as possible (more people means more things that can go wrong), so we’ll have them vote multiple times. As suggested by James O’Keefe’s Project Veritas videos, busing these voters from polling site to polling site is a pretty good way to stuff the ballot, and eco-conscious to boot. Let’s say that each person getting bused around can vote eight times. We could raise or lower this slightly and not make much of a dent in our argument.
FiveThirtyEight shows a 1% gap in projected vote share between Clinton and Trump, which I expect to close to a dead heat, so let’s call them neck-and-neck. If I were going to rig this election, I’d want a 2% margin, meaning 80,000 of our 4 million voters. Since each of our fraudulent voters is going to vote eight times, this means that we need to mobilize 10,000 people. This is going to take some serious organization, especially since each of those 10,000 people needs to be able to produce eight different identities to eight different district officials. The organizers are going to have to create 70,000 identities at minimum, assuming each fraudulent voter uses their own once.
How many buses?
Here’s where things start to get tricky. A school bus has a max capacity of about 50 adult people, give or take. This means that we need around 200 buses. Getting them isn’t necessarily a problem — the Charlotte, NC area has one of the largest busing systems in the US, with over 200,000 students using the system daily, so there’s plenty of buses around. Even more fortunate for our scheme, North Carolina has closed the schools for the day, so the buses will be available.
Logistics
So we’ve got our plan — 10,000 voters will be bused from polling station to polling station, voting once at each station. At first blush this might seem far-fetched, I mean who’s not going to notice a bus full of people rolling up and unloading? As it happens, lots of communities without adequate public transportation organize busing to get people to the polls. It’s convenient and cost effective since by design, people tend not to live too far away from polling locations.
Next we need some routes. This part is hard. We can’t have the same buses going to the same polling places, since that many extra votes in the same locations would raise eyebrows, not to mention bus after bus unloading at the same precinct. With 200 buses making 8 stops each, we’re looking at a total of 1600 stops. There’s fewer than 500 total precincts in the two most populous counties in the state (containing Raleigh and Charlotte), so we’ll likely have to fudge a bit to make sure no precinct gets hit twice.
There’s also a problem with population. We’re adding 80,000 votes to some relatively small geographic areas; remember, each of our fraudulent voters needs to vote eight times in a day, so we can’t go too far afield. Since each bus only holds 50 people, we’re only adding a maximum of 50 votes per precinct, or 400 votes per bus. Luckily, many of the areas we’re looking at (Charlotte, Raleigh, etc.) are fairly populous, so a few hundred votes here and there shouldn’t be too big of a deal.
Solving this route issue is computationally difficult, given all the different constraints and the fact that this is a special case of the Traveling Salesman problem. Still, let’s assume that a savvy programmer works out some nice routes for us such that no precinct gets hit by more than one bus, and that no bus is adding too many votes to a small district.
Getting away with it (?)
All right, we’ve added 80,000 votes to the North Carolina election. We did it in such a way as to ensure that we didn’t dump too many votes in one place, and spread over enough geography that we’re not being too suspicious.
The first issue in getting away with this involves believing that no one will notice an extra 200 buses driving around on election day. Given the recent political climate, I doubt that would pass unnoticed, but whatever.
Next up we need to look at funding. We have zealous voters, so we don’t need to pay for them. According to Bus Bank, you can rent a bus for up to 12 hours for as little as $700 (let’s assume that includes gas), so renting 200 buses will cost $140,000. In order to not get caught, someone would have to coordinate renting these buses in small groups, not as a block. Furthermore, there would have to be separate funding streams for each rental, otherwise we’d get busted as soon as someone follows the money.
Then there’s those 70,000 extra identities we had to create. Even without requiring a government issued ID at each polling place, that takes work and resources, so we’ll have to presume that a third party of some kind is taking care of it. How much does all this cost? I have no idea. But even if this part were free to the masterminds of this scheme, that $140,000 for the buses is still a decent chunk of change. We need high-level actors of some kind involved, there’s just too much organizing to be done.
The rub
So we did it, right? Flipped a state with in-person voter fraud. To do it, we just had to overcome numerous logistical and legal hurdles. Now, our candidate can safely count on North Carolina being in their corner.
And now that victory rests on the hope that at least 10,000 people never breathe a word of what they’ve done to anyone important. Not leading up to election day, not after election day, at least for the next four years assuming our candidate wants to run again. They’d have to keep quiet until the ends of their lives to safeguard the candidate’s legacy.
This is where a lot of NASA conspiracy theories fall down. It strains credulity that a large group of people, at least a few dozen, would keep their mouths shut for decades about something as important as the moon landing, especially when the rewards for coming forward are so great. The idea that the CIA or FBI or someone would assassinate anyone who tried to talk gets into the realm of even lower likelihood; if the government boogeymen couldn’t stop Edward Snowden, how are they going to stop a bunch of random nerds?
Our fraudulent vote scheme has the same problem. We have to believe that either the government can individually police 10,000 individuals for the rest of their lives, or that at the very least there exist 10,000 individuals who would help flip one of the most contentious elections in American history and then just not talk about it until their deaths.
Think about how unlikely that is. Then think about all the caveats we had to make to even get this far: knocking a half million voters off the rolls, getting thousands of extra people to polling places multiple times in a 13-hour voting day, no one noticing the extra buses… It’s just too much.
How to successfully steal an election throughout American history
In the real world, no one counts on in-person voter fraud to steal an election.
Before the secret ballot was instituted in the late 19th century, election day was a raucous affair, with local officials getting folks liquored up to vote for their candidate. This wasn’t really “stealing” an election, George Washington himself bought 160 gallons of liquor for just such a purpose, but boy it sure is easier to sway people when you’re giving out free hooch. Party bosses in places like Chicago and New York put out enforcers to make sure people voted the right way, who went as far as to learn the different noises a mechanical voting machine would make depending on how someone voted.
The best example of voter impersonation in the US was in Brooklyn, NY between 1968 and 1982, but even then the only reason it worked was because local officials were in on the fraud. Furthermore, it’s massively unlikely that any national races were affected, given that the largest number of bogus votes cited was 1,000.
Ballot stuffing, vote-buying, intimidation, vote theft, all of these are embarrassingly common throughout American history. It’s how we got to the voting system we use in 2016. Voter fraud of the type being suggested now, the in-person kind, happens rarely, and even more rarely is it on purpose. It’s just not practical.
