Despite Donald Trump’s recent claims, it’s pretty impossible to rig an election via voter fraud. To have any impact, you’d need a labyrinthine network of local election officials to collude against a candidate and then bamboozle the bipartisan poll watchers tasked with keeping them honest.
But just because Trump’s fever dream of an election day conspiracy is highly (did we mention highly?) unlikely, doesn’t mean that American elections are always—if ever—fair and equitable.
Consider this: In 2012, black voters waited in line twice as long as white voters to cast a ballot. In key swing states like Florida, the largest polling place delays occurred in districts with larger minority populations. Districts with more Spanish speakers also experienced longer lines. And research shows that in 2012, somewhere between 500,000 and 700,000 eligible voters decided not to vote because of problems at their polling places, including wait times.
“It has an economic cost,” says Charles Stewart, a professor of political science at MIT and one of the leading researchers on voting lines and voting technologies, who estimates that some $1 billion in productivity was lost in 2012 due to people waiting in lines. “But whatever the cost is, it lands disproportionately more on some people than others, and that’s unfair.”
So in a way, you could say elections have been rigged all along—just not in the way Trump envisions it and certainly not against the people who are most likely to vote for him this November.
The question is: Why does this kind of rigged system exist? Though much has been made about voter identification laws and the way they undermine the promise of the Voting Rights Act, there’s another insidious problem plaguing American elections, and that is the fact that the machines on which we vote are old and growing older, they’re allocated unevenly, and election officials lack both the funding and the data they need to update them.
A Systemic Imbalance
In his 2012 victory speech, President Obama famously pledged to “fix” the long lines that wound around polling places for hours in states like Florida. Since then researchers and election officials alike have begun probing the root cause of this issue.
Now, one of the leading theories behind why some voters wait minutes to vote and others wait hours is a basic imbalance in the number of voting machines available to different demographics of voters. One comprehensive 2014 study by the Brennan Center for Justice found that in Florida, Maryland, and South Carolina—three states that experienced the longest delays in 2012—districts that had more minority voters also had fewer voting machines per registered voter, and therefore, longer wait times.
Baltimore County, for instance, where 24 percent of the electorate was black in 2012, experienced some of the latest poll closings in Maryland on Election Day. More than 90 percent of its counties also failed to meet Maryland’s mandated minimum for voting machines per registered voter. But that failure wasn’t equally shared across Baltimore County. According to the Brennan Center’s report, only precincts with a higher percentage of black voting age citizens had long delays.
“We were really focused on identifying the issue and showing there was a systemic problem,” says Christopher Famighetti, a voting rights researcher and one of the authors of the study. Famighetti and his team stopped short of prescribing a reason for that imbalance, which would have required analyzing how each district allocated resources.
But Stewart says it’s not a huge leap to say that this gap is just another example of government providing services to black and white communities differently.
“In a neighborhood that has long lines chronically, they probably have parks that are poorly maintained. They probably have schools that are crowded. They probably have slow police response times,” says Stewart. “On average, African American communities and communities of color just don’t get the public services that white communities do.”
What imperils this already vulnerable system even more, however, is the fact that so many of the country’s voting machines and electronic check-in tools are a decade or more old. That means not only are they becoming obsolete, but often, they’re so slow and clunky that they gum up the works.
The last time the federal government invested substantially in new voting technology was after the catastrophic screw up of the 2000 election. The Help America Vote Act, which passed in 2002, sent a $2 billion cash infusion into the states, enabling them to invest in new voting technology to avoid any more hanging chad debacles in the future. More than a decade later, many districts are still using those same replacement machines. As WIRED has written about at length, computers that old pose not just a problem for efficient voting, but also a huge security risk.
“The machines we’re using are computers,” says Famighetti, who also wrote a 2015 report on the country’s aging voting technology. “We don’t expect our laptops to last a decade.”
That study found that this year 43 states will use machines at least 10 years old. The states are aware that’s a problem, too, but in most cases, they lack funding to do anything about it. Of the 31 states who said they want to purchase new voting machines in the next five years, some 22 of them didn’t know where they’d get the money.
What’s more alarming—though not at all surprising—is that the districts that do have the money to invest in new voting technology often have richer citizens. In Virginia, for instance, the Brennan Center found that the median income of the 16 jurisdictions that had recently replaced their equipment was $69,800. In the rest, it was $50,100.
“This has the potential to create a two-tiered voting system,” Famighetti warns.
The older these machines get, the more likely they may be to have irregularities and calibration errors. That’s problematic on its own, it can be even more dangerous when voters are prone to believe the election is rigged. Every error can be used as a data point to prove the deck is stacked against the voter.
In early voting, some such reports are already popping up on social media. In Texas, for instance, voters in several counties reported having their votes “flipped” at the last minute. Some media outlets, including conservative commentator Sean Hannity’s website ran with the story. But election officials in those districts maintain that any issues have been the result of human error dealing with less than intuitive machines.
Famighetti predicts these types of stories could proliferate in the future. “The sort of irregularities that we see in elections, and may be more likely to see due to aging out equipment, will be viewed through this lens,” he says, “and that can undermine the public confidence in the election as a result.”
Where’s the Data?
Researchers have only just begun to look into these problems in earnest. Most of the research we’ve had in the past about wait times is based on imprecise data, cobbled together from county-level reports about what time polls actually close, compared with what time they were supposed to close. It’s a rough approximation, that still probably misses a lot of the precincts where lines pile up in the morning. And even that can’t account for what exactly caused the slowdown. Was it a lack of voting machines? Poll workers? Language barriers? Voter ID laws? Ballot length? Or something as simple as the fact that the polling place was physically smaller than others?
“If the question is what data’s available? The answer is: not very much,” says Michael Herron, a professor of government and quantitative social science at Dartmouth University.
But in recent years, as social media posts about wait times and outsized lines turn into front page headlines, Herron says, that’s starting to change. The onslaught of media attention is forcing election officials to at least acknowledge the problem. Meanwhile, in 2014, the Presidential Commission on Election Administration answered President Obama’s election night call with a list of recommendations about how districts could speed up the voting process, including collecting more data on election night operations.
Since 2008, Stewart has been conducting a national poll called the Survey of the Performance of American Elections, which includes information on voting experience from 200 voters per state. It has now become the basis of other election administration research like the studies conducted by the Brennan Center.
This year, however, Stewart, Herron, and about two dozen other university faculty members across the country are going even farther, deploying their own students to about 1,000 precincts to collect data on how long it takes voters to get from point A to point B, among other things. Herron and other Dartmouth researchers are also working on an app called PollTracker that voters could use to self-report their experiences at the polls.
None of this is easy. Voting is by definition supposed to be a private thing, and voters get mighty anxious about people poking around their precincts. Some states ban that kind of loitering altogether in hopes of eliminating the potential for voter intimidation. But before we start giving into fearmongering about the election being rigged, the country needs a lot more information on the parts of the electoral system that already are.