University of Maryland: Researcher Studies Tesla’s Twitter Bot Boost
Amid Elon Musk’s successful $44 billion bid to buy Twitter this week, there’s no denying the value the social media platform has brought to one of the world’s richest person’s other companies, 240 characters at a time. Tesla’s sky-high valuation—over $1 trillion—is linked to the electric car maker’s maneuvers on Twitter, said a University of Maryland management and entrepreneurship researcher.
In a new working paper recently highlighted in a Los Angeles Times article, Robert H. Smith School of Business Associate Professor David Kirsch identifies a set of non-human accounts known as fanbots, and explores the possibility that these accounts may have influenced the trajectory of the firm by shaping how Tesla is discussed on Twitter.
“The Tesla narrative is extraordinarily powerful,” Kirsch said in the Times. “At a certain point, it does become self-fulfilling.”
Kirsch and his research assistant, Mohsen Chowdhury, are trying to answer whether the tweets from these fanbots—fake accounts created by software that automatically tweet, retweet and like or dislike accounts to amplify some views and mute others—are related to changes in the price of Tesla’s stock.
Kirsch and Chowdhury reviewed Tesla-related tweets from when the company went public in 2010 through the end of 2020 using Botometer, a software program that separates bot accounts from real human ones. They found that one-fifth of the relevant tweets about Tesla were bot-generated—similar to the fraction of tweets about other tech giants like Amazon and Apple. But unlike those other companies, they said, Tesla-related bots seemed to reinforce particular narratives about the companies.
The researchers don’t yet have evidence of a direct link between bot tweets and stock prices, but their research continues. Among their findings so far: Of the 4.2 million tweets containing the “cashtag” $TSLA, 400,000, or 10%, came from the bots identified among this subset of accounts. And of the 157,000 tweets containing the hashtag #TSLA, 23% were from bots.
They tracked 186 fanbots and found that the company’s stock appreciated more than 1.4% in the week immediately after each account was created. Tesla’s stock prices have endured some big ups and downs during that period, but when new bots hit Twitter, stock price increases have followed, the researchers said.
“This isn’t (yet) a causal relationship, but it does raise questions,” Kirsch said. “Tweets alone cannot move a stock; if the tweets are implicated, it must be by activating pathways that get to traders.”
He and Chowdhury are also looking into the timing of tweets and the range of topics within the collections of tweets. One thing they don’t know is whether the Twitter bots’ creators have any direct financial interest in Tesla.
Kirsch and Chowdhury analyzed bots that have been created on behalf of other companies, but found that these seem to spread “generic” pro-market messages. Regardless of the effect on stock prices, bot campaigns are a new form of corporate content distribution; Kirsch calls it “corporate computational propaganda.”
“This computational content may have buffered the Tesla narrative from an emergent group of critics, relieved downward pressure on the Tesla stock price and amplified pro-Tesla sentiment from the time of the firm’s IPO in June 2010 to the end of 2020,” the researchers wrote in their working paper. They say regulators may have to consider the use of Twitter bots in Securities and Exchange Commission disclosure rules.