As COVID-19 blisters across America, state leaders are considering a full transition to voting by mail (VBM) for the upcoming presidential election. To the casual observer this seems timely: With no vaccine in existence, COVID-19 is orders of magnitude more contagious and deadly than the flu, and allowing people to congregate in polling booths in November may add to the severity of its reemergence. Even so, the costs of a purely VBM election — especially in the wake of a crisis that is presently shifting the balance of powers and straining federalism to its limits — seem to outweigh its moderate benefits.
We must come out of this pandemic with the ability to sustain an active and legitimate election process. And that means classifying polling areas as essential services. VBM presents the virus with an indirect opportunity to defect the holy grail of American politics.
The weaknesses of VBM are plenty; via a retinue of logistical complications, it provides leverage for voter fraud, coercion, increased ballot rejection and decreased voter turnout of specific cohorts. Fully transitioning to VBM in six months for a presidential election will only increase these risks.
VBM fraud can be achieved in numerous ways. Consider duplicate voting: In eastern regions with small states, voters need only drive a few hours to register in an additional state and ultimately receive two ballots in the mail. Double voting prospects are statistically ominous; a study published by APSR estimated that roughly one in every 4,000 votes cast in the 2012 presidential election was a duplicate.
If the price of gas come November makes this trip undesirable, voters can instead employ false registration to obtain multiple ballots.
A personal anecdote by journalist Margaret Menge points to the ease of falsely registering in Florida. In 2011, she applied three times using the names Hannah Arendt, Rebecca Bugle and her own, not providing a social security or driver’s license number on any of the applications, as they were not required. Weeks later, she received a voter information card for Hannah Arendt. With elections approaching in 2012, she called the County Supervisor of Elections office and was mailed an absentee ballot. It is curious that the applications for Rebecca Bugle and Margaret Menge did not elicit immediate consequences for Menge or spur an investigation. Thus, a person may send in multiple applications — as many as it takes — to achieve registration.
This reveals a consequence of the Florida VBM system’s heavy reliance on inaccurate voter rolls, which often are not updated as necessary to remove the names of deceased citizens, felons, or non-citizens. But this issue plagues states across the country, further handicapping VBM integrity on a nationwide scale. For example, Colorado, which conducts its elections completely by mail, was among five states who received a warning from the Judicial Watch this January for failure to update voter rolls. Furthermore, the state supervisory administration has no way of determining if a person with a certain birth date exists — as exemplified in Margaret Menge’s account — because they lack a master list of citizens to cross-check.
VBM is inarguably more susceptible to fraud, as both parties seemed to agree in the years preceding Trump. A 2005 commission led by former President Jimmy Carter (D) and James A. Baker III (R) concluded that mail-in ballots “remain the largest source of potential voter fraud.”
VBM is open to coercion and mismanagement by political operatives during ballot harvesting excursions — when intermediaries collect and submit ballots on the voters’ behalf. The elderly and physically disabled are most at risk, both because of vulnerability to intimidation and because they are the largest cohort involved in absentee processes. The bipartisan National Election Defense Coalition recognizes ballot harvesting as a potential disadvantage of VBM in regular circumstances, and a primary reason for avoiding VBM elections. As of 2019, 13 states prohibit ballot harvesting by operatives, and for good reason.
Higher error rates
For a system that is advocated for its convenience, VBM has a notorious record concerning lost votes. Such votes are “lost” in the mailing schematic when:
- Mail ballots requested are greater than mail ballots received
- Mail ballots returned are less than mail ballots received
- Mail ballots counted are less than mail ballots returned
In a study conducted by professor Charles Stewart III of MIT, VBM ballots had an estimated lost-vote rate of approximately 22% in the 2008 presidential election, compared to an overall (in-person and VBM ballots) lost-vote rate of 4% in the 2000 presidential election, in which relatively fewer VBM ballots were cast. Clearly, if 100% of voting had been through mail in 2000, the rate may have been scandalous.
A Heritage Foundation dataset shows that across the U.S., nearly a fifth (19%) of all legally confirmed voter fraud since 1997 involves absentee voting, which is essentially the voluntary version of VBM. This is important because in the 2016 general election, mail ballots accounted for less than 24% of the total ballots cast. Combining the prospective lost vote rate based on the above calculations and the legally confirmed voter fraud rate, a move to 100% VBM in November would undoubtedly result in less integrity per vote than in-person voting.
These estimations, while not perfect, are representative and nontrivial enough to raise serious worries about a purely VBM presidential election. The schematic by itself testifies to the multi-step, high-risk pipeline for a mail ballot.
Lower turnout of minorities; higher turnout of partisans
Many VBM supporters assert that because voters need not leave their domicile, overall turnout is increased. Research on this topic is inconclusive, and is representative of individual states only — notably Oregon and Colorado — who have had decades of experience with VBM en masse. The increased turnout claim is further weakened not only because the combined effects of moderately increased turnout, increased fraud and vote loss cancel at best, but because VBM has been shown to favor affluent cohorts and partisans.
With VBM, a postal worker is given a high responsibility, and will encounter difficulty in impoverished, minority-dominated urban communities where the population is more nomadic. As Teresa James and Micheal Slater of Project Vote explain, “First class mail, as its name might imply, does not treat everyone equally. In fact, it discriminates against low-income communities and dense urban areas where residents move more frequently and not every adult shares the same family name. This bias is codified in the U.S.P.S. Domestic Mail Manual, which provides that if ‘the addressee of certain mail is unknown to the delivery employee, the mail may be withheld pending identification of the claimant.’”
Focusing again on Oregon, MIT political scientist Adam Berinsky argues that “VBM in Oregon accentuated the stratification of the electorate … VBM mobilized those already predisposed to vote — those individuals who are … registered partisans — to turn out at higher rates than before.” Therefore, non-habitual, traditional voters, who with a full transition to VBM face a novel set of election materials, will likely be less motivated to participate.
To determine the appropriate language for each household in 2020, the U.S. Census Bureau mailed multilingual inserts, with invitations in 12 non-English languages to verify their preferred language by phone or online. It is easy to imagine yet another batch of “lost” mail ballots this November due to oblivious practices, mail theft, or failure to complete a similar procedure.
Additionally, the mail-ballot pipeline poses an innate risk far greater than the algorithmic hardware within voting terminals. Skeptics of this claim are tasked with proving that postal workers, political operatives, election officials, scanners, tabulating machines and personnel are more reliable than a script of code.
If in 2008 in-person voters were told by their respective supervisory committees that their terminal input had a 10% chance of defecting, there would have been national outcry. Yet VBM has been shown to pose a greater threat to legitimacy, yet there is little concern among mainstream media.
For those who point to health concerns, the risks of venturing to poll booths will not be sufficient this autumn to justify a gamble with VBM that endangers the national election process. Many individuals in the U.S. venture to essential institutions every day, accepting the risks involved because survival mandates that they do. They wear protective gear and complete their various tasks. Highly frequented areas are regularly cleaned.
This is an example of practical adaptation to a crisis, because the fundamental mechanisms by which essential services are offered were not changed. Instead of utilizing Congress’s $400 million aid package to completely overhaul the traditional voting method, the arguably essential services polling areas provide should be practically revised by states to fit the circumstances.
Excluding those with health conditions and the elderly, a legitimate 2020 election requires that able citizens present themselves to vote as they regularly do. This comes with unavoidable risk, and must be taken when the stakes concern the presidency — that may be, as in the past, decided by mere hundreds of votes. People should undoubtedly be vigorously admonished to exercise hygienic discretion in regard for others’ health. Sad is the day, however, when governments forgo their trust in their constituents to implement even these simple practices, on one of the few occasions they can directly influence the course of the nation.
A pure VBM election would cause controversy beyond imagination, weakening the fabric of America’s voting system and launching fruitless investigations at the cost of taxpayer dollars. This is because Americans do not have the confidence in VBM that is requisite for its legitimacy. Stanford computer science professor David Dill put it aptly: “Ultimately, [the] debate is about public confidence in our system. The real job of an election is not to convince the winners that they won, which is often easy, but to convince the losers that they lost.” Americans inherently do not trust VBM — the 2008 Survey of the Performance of American Elections found that only 16% of Americans supported laws that would require all votes be cast by mail. Twelve years of full-fledged technological advancement later, it is doubtful this opposition — controlling for virus concerns — has moderated.
More from Dill: “the public must know that [the] results are accurate. … A helpful analogy is: suppose the voters dictated their votes privately and anonymously to human scribes, and that the voters were prevented from inspecting the work of the scribes. Few would accept such a system on simple common sense grounds.”
Indeed, this reasoning should hold for the approaching election and all that follow.
A previous version of this article stated a finding of the APSR paper in stronger terms than the paper used itself. The article has been updated to better represent the APSR study cited. The Daily regrets this error.
Contact Jeeven Larson at jlarson7 ‘at’ stanford.edu.
The Daily is committed to publishing a diversity of op-eds and letters to the editor. We’d love to hear your thoughts. Email letters to the editor to [email protected] and op-ed submissions to [email protected]