Are bots propping up Tesla’s stock? Two researchers share why there is reason to investigate

In an interview with Moneycontrol, they talk about their paper and the use of social-media posts to build a narrative to boost stock prices of publicly traded companies, and the extent to which market regulators can act against this.

What led you to study social-media posts’ influence on Tesla’s price movements? Prof: Kirsch: I have a long-standing interest in the electrification of transportation and asset bubbles. (Kirsch has authored two books Bubbles and Crashes: The Boom and Bust of Technological Innovation (2019) and The Electric Vehicle and the Burden of History (2000)). Both came together in Tesla, which I wrote in my second book was overvalued according to fundamentals and that it represented a bubble. A little over a year ago, we (Kirsch and Chowdhury) were talking about it… this is an area of common interest, and we thought we would look into the social-media activity around it. We noticed two movements in the stock prices, one was pro-Tesla and the other was anti-Tesla, and there seemed to be hashtag activism around those two movements.

Which markets/sectors do you see this behaviour–of social media being used to influence stock prices or valuation–happening? Kirsch: This seems to be happening in more emergent ecosystems, such as the electric vehicles’ space or in the cloud computing space. Where social-media influencers have a part to play and where there isn’t a clear industry structure, or it isn’t clear who the company is competing with or there aren’t obvious operating metrics (for example, sales per square foot in retail). For these, you need a narrative to make sense of what the firm was doing and to shape people’s expectations about the future performance or future value of the company.

Chowdhury: It is interesting how in a known industry, a narrative would only go so far. On the other hand, in an emerging industry, particularly with a lot of startups, narratives go a long way. If you think about accelerators, they train startup founders to sell narratives and not necessarily the product or service.

With startups, they are largely privately held. In the case of Tesla and other publicly listed companies, and this narrative building, are market regulators equipped to deal with this? Kirsch: Well, they have tried, I think. The Securities and Exchange Commission (SEC) in the United States exists to ensure transparency and accuracy of information…But, everyone’s always trying to outsmart the regulator and the question is if the regulator is getting smarter fast enough. This is a bit outside of my expertise, but I think this is one area (new channels of information dissemination) where market regulators need to be focussing their energies on. For example, in 2018, when SEC fined Tesla and Musk $20 million each (for posting “funding secured” to take the company private at $420 a share, which caused volatility in the stock for weeks) but if he made say a few hundred billions in excess market value (because of that tweet), then the fine is just a parking ticket.

So, market regulators have the authority to act and they have the resources to act as well… but sometimes, depending on the political machinations and depending on which party is in power, the regulator may not be encouraged to act (against such behaviour). The regulator then swings between acting vigilant and being more tolerant.
When we were interviewing people for this working paper, one of them was someone from TSLAQ (a loose grouping of Tesla sceptics), and this person saw TSLAQ as a sort of open-source SEC. They believed that the market regulator wasn’t doing enough to rein in market abuses by the company. So, they viewed their spreading critical information about the company as doing a job that the regulator wasn’t doing.

Can you actually measure the rise in stock price from such storytelling, to what degree or level the price changed following these posts? Kirsch: Mohsen and I have been joking that that is the trillion-dollar question. You know, if Tesla was destined to be a failed startup but ended up being a trillion-dollar company because of social media. As of now, ours is a working paper. But, we believe that we will be able to show some effect of these tweets on the Tesla stock. As of now, we can see that the stock and the tweets sort of move in the same direction. So we have correlation but we are still trying to nail down causation.

What are the challenges to proving that there is causation?

Kirsch: Stock markets process a lot of information. So it’s very hard to cleanly attribute an x to y. There are theoretical and empirical challenges. Theoretical in the sense that we need to understand to what extent these posts are influencing people’s buying behaviour and so on… we need to understand the social process these fan bots (set off). Empirical because we need a way to measure this (influence).

When did you find the bots most engaged? Chowdhury: We just observed a few correlations. The bots were most active on days the stock was under pressure or had a negative market-adjusted return. The bots were less active when the stock had a positive market-adjusted return, they were less active when something good happened. That is what led us to believe that the bots may be buffering the stock on days the narrative was under pressure. I must add that these are very simple, early-stage correlations and the analysis is ongoing.

You have said that the bigger question is who’s actually running these bots. From what you have seen, does this operation look like something that would need a massive investment? Kirsch: There are about 186 bots we have noticed. So this does not require a lot of investment or management. This isn’t a secret underground office with thousands of desks. When the Chinese Communist Party was trying to attack the tennis player Peng Shuai and her credibility in 2021-2022, they were said to be using about 97 bots. In that context, Tesla having around 186 bots working for the last ten years looks like a pretty significant campaign… bear in mind that 186 bots can generate content as much as thousands or even millions of human users. The bots can post something every three hours, forever. But it is still cheaper than an annual advertising budget.

During your work on this paper, have you noticed market regulators in any part of the world tackle this effectively?

Chowdhury: Nothing that I’ve seen, but increasingly there has been talk about having a warning label with artificially created social-media posts. But I haven’t seen this enforced anywhere as a regulation.