But today, the availability of all this data about elections leads to an increasingly familiar problem: cherry-picking. Thanks to the internet, the pattern-finding obsession depicted by Russell Crowe in A Beautiful Mind has become a kind of pastime for conspiracy-inclined politics buffs: Grab a quote from Whatever.gov, match it to a quote from a Biden speech, share it on X, and you’ll get a flurry of retweets.
Or, if you’re good with numbers and computers, dig into the data sets related to the 2020 election (county-level voting returns, census tallies, etc.) and uncover all the voter fraud Donald Trump claims. You could even make a movie out of it and make a ton of money. It’s been a buyer’s market for nearly a decade, and it shows no signs of slowing down.
Take Douglas Frank, for example. After the 2020 election, Frank became a minor celebrity in Trump circles for his mathematical assessment of how widespread voter fraud was. His analysis wasn’t particularly robust; he used the results from a few counties in a state to extrapolate the data for the rest of the counties, and suggested that the accuracy of his estimates evidenced some process of manipulation. But he became a minor celebrity anyway, giving presentations before Trump speeches and serving as MyPillow CEO Mike Lindell’s sidekick on the voter fraud event circuit.
The problem here, as with all cherry-picking theories of fraud, is that the availability of information far outpaces our understanding of the systems that produce it. Frank thought he found evidence of fraud by poking numbers into a spreadsheet, but he made no attempt to explain how such fraud was possible in multiple states and counties with different voting and counting processes. Frank and others like him portrayed themselves as Sherlock Holmes discovering fingerprints at a crime scene, when in reality they were just piecing together tiny arcs of lines from different places on the surface of the moon.
Those who fervently believe that voter fraud is widespread are typically unfamiliar with existing checks to prevent it, in part because those checks are often overt or hidden. For example, voter registration itself is a key component. Matching voters’ names and addresses makes it much harder to commit fraud. Fraud is uncovered in every election, and a small number of criminal prosecutions have been made for double voting and other illegal activities. The numbers are small in part because there is little attention paid to voter fraud. Would you commit a federal crime to give your state one more vote for your preferred presidential candidate? Sometimes arrests are made in connection with larger conspiracies, which involve more people and therefore are more likely to be caught.
In other words, there is evidence that the system works well as is, but no evidence that it won’t work. Which brings us to Elon Musk.
On Wednesday, The Wall Street Journal reported that Musk, the Tesla CEO and X owner, has been growing increasingly close with Trump. This isn’t a huge surprise — Musk’s pro-Trump politics (and Trump-like personality) have been clear for some time — but one new detail stood out.
“Musk and [investor Nelson] “Mr. Peltz has told acquaintances he is working on a data-driven project to ensure votes are counted fairly, echoing Mr. Trump’s accusations of widespread fraud in the 2020 election,” the Post reported. “The two men briefed Mr. Trump on the effort during a meeting in March, according to a person familiar with the discussions.”
It’s a classic Musk idea: take an existing product and improve it with his own engineering talent. This has worked quite well for Tesla, but not so well with the company’s recent acquisition of Twitter.
Unfortunately, the plan to layer data analytics on elections to root out fraud looks more like Twitter’s efforts than it does Tesla’s. First, it’s clearly rooted in Musk’s increasingly surrealist politics. He’s not trying to prevent voter fraud because it’s rampant; he’s doing it because the bubble in which he operates claims that voter fraud is rampant, just as the same people in that same bubble have inaccurately claimed that Twitter is unfairly silencing right-wing voices. Second, the idea isn’t particularly revolutionary.
This approach has been tried before, and not just by people like Douglas Franks. Trump himself created a presidential commission focused on uncovering widespread fraud after winning the 2016 election. The commission failed without uncovering any widespread fraud. The Trump campaign hired researchers to prove fraud after the 2020 election. They found nothing. The Trump campaign kept it a secret.
Trump was not alone. The George W. Bush administration also tried to root out voter fraud but found no evidence of significant levels of it. Trump allies such as former Kansas Secretary of State Kris Kobach (who also served on Trump’s committee) tried to prove widespread voter fraud but were unable to do so.
But as always, Musk seems to think he can succeed where others have failed. The problem isn’t that the system works, that there are checks in place to prevent fraud, that fraud isn’t rampant. Rather, it’s that we haven’t applied a sufficient level of genius to the problem, and once we do, reality becomes clear like a corpse flower. In theory, it seems like all we need is someone good with numbers.
Douglas Frank thought so too.
