A documentary directed by conservative commentator Dinesh D’Souza claims it can prove widespread fraud was carried out during the 2020 presidential election in the United States. Reuters Fact Check examined the main claims presented in the film and did not find any concrete evidence definitively showing proof of fraud.

The 90-minute film "2000 Mules" sees D’Souza team up with True the Vote, a Texas-based nonprofit that describes itself as protecting election integrity (www.truethevote.org/about/), to investigate alleged voter fraud in Arizona, Georgia, Michigan, Pennsylvania and Wisconsin.

All five of the listed states were swing states in 2020 that ultimately backed Joe Biden for president - and were later central to baseless speculations of fraud.

Reuters has covered this topic extensively (here) and (here), as well as in fact checks (here), (here) and (here).

D’Souza’s documentary says Biden victories in swing states could be thanks to 2,000 people – or “mules” – who were hired by unnamed nonprofits - dubbed “stash houses” - to conduct “ballot trafficking”, i.e.: stuffing numerous drop boxes with potentially fake absentee ballots.

It also alleges that the so-called “mules” were paid $10 for every fake ballot they submitted.

D’Souza did not respond to multiple requests for comment from Reuters.

THE METHODOLOGY

D’Souza and True the Vote analyzed surveillance footage of drop boxes mostly from Georgia, as well as “some” from Arizona, along with “geotracking” data purchased from unnamed brokers.

The “geotracking” data was gleaned from cellphone apps pinpointing device location and movements between Oct. 1, 2020, and election day, Nov. 3, for Arizona, Michigan, Pennsylvania and Wisconsin, according to the documentary. Data for Georgia stretched until January, when there was a runoff vote.

The documentary alleges that by tracking phone locations to the addresses of five alleged “stash house” nonprofits and 10 or more drop boxes, the “mules” were identified.

There were 242 people in Atlanta, Georgia, who fitted the bill; 200+ in Arizona; 100 in Milwaukee, Wisconsin; 500+ in Michigan, and 1,000+ mules in Philadelphia, Pennsylvania – totaling over 2,000 “mules”.

Viewers were then shown multiple surveillance footage clips of different people at drop boxes, which the documentary said it had identified as some of the ballot traffickers carrying out their crimes.

GEOTRACKING

Multiple concerns were raised by experts speaking to Reuters about the “geotracking” portion of the documentary. It was unclear whether the same test was applied anywhere other than the swing states in question (to prove a unique phenomenon had happened), along with data validity, accuracy, and discussion about other possibilities that could explain the findings.

“The entirety of the claim rests on cell phone location data, which doesn’t remotely show that people were actually using the drop boxes (it doesn’t have the granularity to show that, as opposed to just walking or even driving by),” said Kenneth R Mayer, Professor of Political Science at the University of Wisconsin-Madison, who spoke to Reuters via email.

According to True the Vote founder Catherine Engelbrecht, who spoke in the documentary, the dataset had been validated because it was used by the organization to solve two murder cases that were “ebbing on cold case status”.

Only one murder case was detailed as an example in the documentary – that of eight-year-old Secoriea Turner on July 4, 2020, in Atlanta – and which authorities told NPR was solved without anything to do with Engelbrecht (here).

D’Souza, meanwhile, claimed without offering evidence that the dataset had the “reliability of a fingerprint”, expanding in a later podcast interview that it was accurate to between “12 and 18 inches” (here).

Experts speaking to Reuters disagreed.

“I have never heard that geotracking using cell phones could have errors as low as 12-18 inches,” said Chen Qian, Associate Professor at the Department of Computer Science and Engineering at University of California, Santa Cruz. “This range is way below the ranges reported by scientists and engineers.

“A research paper written by AT&T and Purdue University researchers in 2020 predicted that the average location error of 5G networks would vary from 2 meters to >10 meters,” said Qian. “Note their results are simulated results in ideal settings, used for predictions. They are not real experiments, because 5G has not been available everywhere. In real environments the errors would be larger.” (here)

Moreover, drop boxes tend to be in high-traffic areas such as public libraries, shopping malls, municipal buildings, or schools. For example, a map of drop boxes in the five metro Atlanta counties of Fulton, DeKalb, Cobb, Gwinnett and Clayton shows that all are clustered in busy locations likely to have high cell phone activity (here).

“My local drop box is in my public library, a location I pass probably 20 or 30 times a week,” said Paul Gronke, Director of the Elections and Voting Information Center at Reed College in Oregon (evic.reed.edu/). “Did I deposit 20 ballots or is my drop box on a heavily trafficked street? You tell me.”

True the Vote said in the documentary it had ruled out people where it believed their “pattern of life” outside the election period involved travelling to nonprofits and drop box locations. They did not offer information on how they did this or who these people were.

However, Barry C Burden, Director of the Elections Research Centre at the University of Wisconsin- Madison (elections.wisc.edu/), told Reuters via email that there were still acceptable reasons for observed heightened activity during election periods.

He said: “Some of the individuals tracked might even have been election workers checking on or emptying the drop boxes, so it would be a sign of vigilance by election officials rather than nefarious behavior.”

SURVEILLANCE VIDEO

To corroborate the geotracking dataset, True the Vote said it had compared it with surveillance video covering locations of some of the drop boxes.

They claimed to have access to 4 million minutes of footage, which was mostly from Georgia. The documentary makers said some of the surveillance cameras were turned off in Arizona and that there was no footage from Wisconsin. No information was provided about surveillance footage from Michigan or Pennsylvania.

Reuters was unable to examine these alleged minutes of footage, but the videoclips presented in the documentary alone do not provide proof of fraud.

When Reuters asked Engelbrecht via email how “mules” identified via geotracking data were then matched to surveillance footage, she responded: “Matches are made by comparing the location and time stamp of the video to the location and the time stamp of the individual device.”

The documentary shows several surveillance clips said to reveal “mules” stuffing fake ballots in the drop boxes.

In one clip, a couple of ballots appear to drop to the floor as one man goes to post; the documentary makers suggest this is suspicious, as well as the man allegedly posting the ballots late at night.

In another, a woman wearing a face mask and gloves is seen posting a ballot before turning to place her gloves in a nearby bin. It is claimed in the documentary that she is a “mule” because she was wearing gloves (to hide her fingerprints) and did not look at the bin, so must have had prior knowledge of it being there. The documentary makers did not appear to consider the possibility that the woman was wearing gloves, along with her face mask, as a personal protective measure against COVID-19.

The unidentified woman was also said to have visited “dozens and dozens” of drop boxes; however, no other clips of her, nor any further evidence, were shown.

Two other clips show men in separate locations taking photos of themselves posting their ballot, which the documentary makers allege was to provide evidence of the job done so the “mules” can get paid.

In a Fox News interview, Engelbrecht claimed the average number of visits by a "mule" to a drop box was 38 (here). Yet none of the surveillance videos showed the same person more than once.

BALLOT HARVESTING VS FRAUDULENT BALLOTS

Some of the people in Georgia who were presented in surveillance footage as so-called “mules” were seen posting more than one ballot at once, which the documentary makers suggested was proof of voter fraud.

But this doesn’t necessarily constitute fraud. Ballot harvesting, the posting of completed ballots on behalf of a third party, is legal in several states, including Georgia (n-thhere).

“Some of the so-called “mules” might have been legitimate family members putting in ballots in Georgia,” said Theodore Allen, Associate Professor at Ohio State University, specializing in the administering of elections.

“Many people need encouragement to vote and offering to collect and bring to ballot boxes is, in many states, a legal and legitimate way to increase voter participation which is often low.”

Reuters also spoke to M.V. (Trey) Hood III, Professor in the Department of Political Science at the University of Georgia, to understand the legality around dropping off multiple ballots. He said that issues of ballot harvesting and fraud have been conflated, and there are protections in place in Georgia to prevent the posting of fraudulent ballots.

“[In Georgia] In 2020 we were using signature verification to verify absentee ballots, so there were safety mechanisms in place to ensure it wasn’t a fraudulent ballot.”

He added: “I haven’t seen any hard evidence being offered up that these ballots were fraudulent.”

Reuters looked into a video posted in May claiming to show one such “mule”. Pennsylvania county authorities debunked the “evidence” by confirming it showed a designated agent dropping ballots off on behalf of individuals who are unable to (here).

Ultimately, ‘2000 Mules’ speculates that the so-called ballot-traffickers were dropping off fraudulent ballots – but the film does not prove this. The ‘faked’ ballots were never opened and inspected, nor were the suspected “mules” on surveillance questioned, aside from one anonymous “informant” who says she never saw inside the supposed fraudulent ballots.

COULD BALLOTS BE FORGED?

The documentary claims its investigation reveals the potential existence of 380,000 overall fraudulent ballots. And, it claims, if all of these contained falsified votes for Joe Biden, the revelation is significant enough to have blocked a win by Donald Trump.

Listing all of the steps needed to falsify a ballot, Gronke told Reuters: “1) You need a falsified ballot with a unique bar code, printed on special paper, and a special envelope. If the claim is that you’ve somehow obtained 400,000 original ballots without the elections officials or voters knowing, how precisely did you do this?

“2) You need to successfully forge the voter’s signature. 3) You need to deposit the envelope and have it validated by a local official.

“Congratulations! Besides committing a felony, you have now cast ONE fraudulent ballot. Now you need to figure out how to do that hundreds of thousands of times, in different jurisdictions, with different ballot styles and different voting materials.”

Gronke’s sentiments were echoed by Christopher B. Mann, Associate Professor of Political Science at Skidmore College, who told Reuters: “If there are 400,000 people who had their ballot collected and returned for nefarious reasons, there should be significant numbers of people willing to tell their story. It is hard for two people to keep a secret. Asserting that 400,000 people are keeping a secret is beyond credibility.”

Reuters has previously explored the safeguards in place that make forging ballots on behalf of others difficult (here), (here).

2020 ELECTION FOUND SAFE AND SECURE

False claims pedaled by former U.S. President Donald Trump and his followers blaming widespread voting fraud for the 2020 election results have been rejected by courts, state governments and members of his own former administration (here).

U.S. election security officials have said the election was “the most secure in American history” (here), (here).

Former U.S. Attorney General William Barr, the nation's top law enforcement official under Trump, said on Dec. 1, 2020, that he had not seen any evidence of fraud that would have changed the election results (here).

Furthermore, more than 50 lawsuits brought by Trump or his allies alleging election fraud or other irregularities were dismissed by state and federal judges (here).

VERDICT

The documentary “2000 Mules” does not provide any concrete, verifiable evidence of widespread voter fraud in the 2020 election. Technology and election integrity experts consulted by Reuters also did not find the geolocation, surveillance or any other information presented showed plausible evidence of fraud.

This article was produced by the Reuters Fact Check team. Read more about our fact-checking work here .