Is there Fraud Impact of Increased Absentee Voting?

E. John Sebes
Co-Founder, Chief Technology Officer

Spoiler Alert: Risk assessment indicates “Not Much.” Let’s explore why.

Photo courtesy: Frederick County Elections Director, Stuart Harvey

With our first presidential election during a pandemic less than 80 days away, it’s become clear that far more people than usual will be voting at home in November. As our analysis shows many states make it easy, some are making it at least temporarily easier, and at this writing, a couple are upholding pre-COVID home voting hassles as hoops for a voter to jump through. Absentee or by-mail voting is sure to far exceed the 26% share it held in 2016, thus, there’s persistent discussion in the media, encouraged by certain politicians and influencers, about the risk of increased fraud from increased by-mail (or “at home”) voting.

As I set out to write this article (early July) I had yet to see a professional threat and risk assessment for increased by-mail voting. Of course, to a degree that changed in late July when CISA published this fine piece of work, which is a different, albeit detailed look at risks. [1] My focus here, however, is on fraud risks, based on decades of experience in risk assessment, combined with over a decade of intense immersion in the nuts and bolts of how U.S. elections work. But first, here’s the top-line:

I don’t see any reason to worry that expanded at-home voting will increase fraud to a level where the integrity of the election would be in danger.

Let’s start by addressing what “voting fraud” means, because it’s understood in very different ways by different people. For purposes of discussion here, I separate “fraud” into two (2) parts:

1. Bad deeds by election officials (in fact, the Department of Justice catalogs a dozen); and
2. Bad deeds by voters or people that they interact with (in other words, intentional corruption of the electoral process by the voter)

The short of it is, it’s extremely rare that fraud is committed by the voter. It is equally rare fraud is committed at the hands of election officials, campaign workers, advocacy groups or even the candidates themselves.

Fraud committed by election officials is real but rare (for example, cases in Kentucky in 2010 and Pennsylvania in 2020). The methods for election officials (EOs) to game the system do vary between in-person voting and at-home voting, but there’s no evidence that one is more effective than another.

Mischief can include ballot box stuffing, forging absentee ballots, falsifying tabulation, or even conspiring to prevent people from voting by multiple means including intimidation.

With that in mind, let’s focus on fraud related to voters. There are three kinds of voting fraud that you can read about in mainstream news and opinion: in-person voting fraud, ballot fraud, and bribery and coercion in at-home voting.

1. In Person Voting Fraud

In-person voting frauds include a voter voting more than once, or a fraudster impersonating a real voter to steal their vote. Every academic study, as well as law-enforcement history, show that this is rare, but still occasionally found and prosecuted. This infrequency is understandable, because a calculating fraudster aiming to avoid detection will not be committing crimes in person in a voting place, when there are less personally risky alternatives, namely, abusing absentee voting. With fewer people voting in person in 2020, I don’t think we need to worry about an increase in in-person voter fraud.

2. Ballot Fraud, and How it Relates to “Harvesting”

Ballot fraud is committed when someone uses or tampers with another voter’s ballot (for purposes here, “absentee ballots”). Even though there will be a lot more ballots mailed to American homes this year than ever before, let me explain why the risk of result-altering ballot fraud remains low.

Elections officials defend against absentee ballot fraud by various means, most notably signature matching, where they compare what’s signed on the envelope with a reference from sources such as the original voter registration card or a DMV database. That’s why a commonly purported form of ballot fraud, where someone intercepts and fills in another person’s ballot, is apocryphal — election officials really do scrutinize and sometimes reject ballots for signature mismatch. In other words, thanks to signature matching, we can trust that the voter really did mark and send their own ballot. (But see “Bribery and Coercion” below for a form of fraud that bypasses this safeguard.)

Some variations of ballot fraud are real, in the form (often unintentional) of what is pejoratively known as “ballot harvesting.” However, let’s be legally clear: most states either explicitly allow, (or at least do not prohibit) someone other than the voter to convey (deliver) a ballot from the voter to either the mail or directly to the election office — as with many topics related to voting, the rules vary widely between states. If you’re in Arkansas you may be committing an illegal act by taking your elderly grandparent’s ballot to a mailbox, whereas in Montana voters recently authorized a wide class of people including “acquaintances” to deliver someone else’s ballot. So, this is a state-dependent matter, however with a few exceptions transferring a signed, sealed ballot is not illegal, nefarious, or tantamount to a fraudulent act (again, bear in mind the signed, sealed envelope to be inspected and matched by the election officials).

But the real issue here is organized conspiracies to influence election results, whether by altering or destroying multiple voters’ ballots. As we saw in 2018 in Bladen County, North Carolina — a state where it’s a felony for anyone but a near relative or guardian to return another person’s ballot — it is possible to run such a conspiracy, and possibly affect an election result, but it’s hard to do without getting caught (but that’s best explained in an entirely separate article). The NC State Board of Elections rightly decided to re-run the 2018 election for 9th Congressional district, when it looked like the margin was less than 1,000, and the number of targeted ballots was rising toward 100 as an investigation progressed.

For organized “harvesting” efforts to successfully flip an election result, it needs to be very close, where approximately 100 or fewer targeted ballots might affect an election outcome. It doesn’t matter whether a state has 6% or 70% at-home voting: there are plenty of voters even at 6% to support a ballot harvesting scheme that would switch a fraction of a percent of votes in a very close contest. Importantly then: that is why an increase in at-home voting does not significantly increase the risk of a close election flipped by ballot fraud.

3. Bribery and Coercion

The third and final type of fraud risk to evaluate for by-mail or at-home voting is bribery and coercion. This is where a legitimate voter casts their one legitimate ballot, but with ballot choices that they made not out of their real choice, but because someone paid or made them feel compelled to vote a certain way.

Bribery is certainly possible, but organized bribery schemes haven’t been documented, despite persistent urban legends about fraudsters buying blank ballot kits at large public events. Coercion, especially so called “kitchen table coercion,” where one member of a household monitors another’s bubble-filling to ensure they vote the “right way,” is essentially an unavoidable risk in at-home voting, and we frankly do not know how often this happens.

Wait: So, increased at-home voting will increase opportunities for coercion, right? Maybe, but again, this is a risk assessment so let’s focus on the threat: an election result flipped by fraud.

There’s no evidence that single instances of at-home coercion or other forms of inappropriate influence have any partisan or other voting pattern. With more at-home voting, we might have more instances of a single coercer looking over the shoulders of household members as they mark their ballot. But more of that, by itself, doesn’t necessarily change an election result one way or another, certainly no more than myriad other factors that influence the electorate’s opinions or ability to successfully cast a vote. To change an election result this way, you need a lot of household (or similar) coercion, organized in favor of a particular candidate or party — in other words, a criminal conspiracy, perpetrated by organized people working for a political goal, which a state has a compelling interest in uncovering and addressing; that is, it can and will be prosecuted to the full extent of the law.

Organized coercion is a real threat, regardless of whether it is rare, so let’s consider whether this threat is more significant with more voting at home. A careful threat analysis gives us a resounding “No.” Let’s take it by cases.

1. In fully-vote-at-home states, nearly every voter is vulnerable to organized coercion because nearly every voter can vote at home.
2. In states where every voter has the opportunity to vote at home, sometimes called “no-excuse absentee voting,” each voter has to request an absentee ballot. Every one of those voters is vulnerable to organized coercion to be compelled or bribed to first request and get an absentee ballot, and then vote that ballot per the coercer’s directions.
3. That’s even true in states where an absentee ballot request requires an excuse, so that the only voters eligible for an absentee ballot are those that are (for example), disabled or out of the country. If we’re considering a criminal conspiracy, we have to admit that in a two-step process a coercer can target all voters, whether their statement of excuse is true, or is a coerced falsehood.

It doesn’t matter if the fraud occurs via a one-step voting process or a two-step process where the first step is making an absentee ballot request. An organized conspiracy could target any voter, in any state, regardless of the percentage of voters who usually vote at home. Accordingly, an increase in at-home voting does not increase the risk of a successful conspiracy to organize coercion to the benefit of one candidate or party.

Conclusion

Should we be concerned that increased at-home voting will lead to more fraud? I think that any organized election-related fraud conspiracy can only operate undetected at a small scale, whether it is “ballot harvesting” or organized at-home coercion.

We should be concerned that fraudsters can nibble at the edges of very close elections. But this is not new, and has been a concern for some time. We know from a long history of U.S. elections that very close contests, with small fractions of a percent difference, can report the wrong winner as a result of a large number of factors malicious or accidental — Florida in 2000 most quickly comes to mind.

But the important result from this threat and risk analysis is that the various types of “voting fraud” threats have been with us for many years, operate at a small scale, and can target an entire electorate. As a result, a shifting balance between voting at home versus in person voting does not increase the actual risk that a very close election will be “flipped” by fraudsters operating undetected.

__________

End Note

1. The detailed CISA by-mail voting risk assessment published a couple of weeks ago focuses on infrastructure issues. That CISA risk assessment is designed to assess the risks to the by-mail voting systems, processes, and infrastructure. It’s a companion to CISA’s Election Infrastructure Cyber Risk Assessment and the Risk Management for Electronic Ballot Delivery, Marking, and Return that CISA jointly released in May 2020 with the Election Assistance Commission (EAC), the National Institute of Standards and Technology (NIST), and the Federal Bureau of Investigation (FBI). This CISA risk assessment only examines the specific risks to the election infrastructure and operations that are associated with by-mail voting. That noted, it’s certainly worth a read for a different and equally important risk assessment.