With the rise in recent years of investment advice proffered by finfluencers, Professor Schürhoff became interested in how finfluencers on social media platforms shape market attention. When the opportunity arose to have access to very granular data on the behavior of finfluencers, Professor Schürhoff and his coauthors wanted to use it to determine whether or not finfluencers' advice carries information and how informative it is.

"There was disagreement in the literature about whether financial advice through social media provides a signal or whether it is just noise. Many believed the latter—that it is just opinions and that those opinions do not carry much information content. We used high-dimensional machine learning techniques to find out."

Not a Lot of Wisdom among the Finfluencers or in the Crowd
Professor Schürhoff and coauthors examined 72 million tweets from over 29'000 finfluencers on StockTwits (a kind of Twitter for finance) and measured the investment performance associated with following each finfluencer's advice. In an efficient market, social media users should gradually identify and follow those who provide genuinely valuable information. Over time, skilled finfluencers would attract larger audiences, while less capable ones would lose relevance and exit the space. This process would ensure that high-quality insights dominate and contribute positively to market efficiency. Against this backdrop, the researchers investigated the skill composition of finfluencers and classified them into three types: skilled (those that generate positive alpha), unskilled (those that generate zero alpha), and antiskilled (those that generate negative alpha). Distinguishing skill from luck is central to this classification, as even poor forecasters may occasionally be right.

The researchers found that while the average skill level of the finfluencers was about zero, the majority were actually antiskilled, giving advice that would result in average monthly returns 2.3 percent behind the market. You would think that the poor investment advice of this antiskilled majority would drive followers away from their channels, perhaps toward the 28 percent found to be skilled finfluencers. But actually, the researchers saw that skilled finfluencers have fewer followers than the antiskilled and unskilled. What seemed to matter most for popularity was frequent, entertaining engagement, even if that engagement ended up being overly optimistic and near price peaks.

These results are directly relevant for investor protection and market integrity. The researchers' findings highlight the need for policy interventions that could redirect attention toward higher-quality finfluencers by improving transparency, active quality tracking, and verification. Professor Schürhoff and coauthors are continuing work with the data, and are now researching the interrelation between the information on social media and how asset prices are formed.

More Information

• Finfluencers, Swiss Finance Institute Research Paper No. 23-30 (with A. Kakhbod, S.M. Kazempour, and D. Livdan) 

SFI Prof. Norman Schürhoff (UNIL)

Norman Schürhoff is Professor of Finance at the University of Lausanne and holds an SFI Senior Chair. His work has been published in the top academic journals in finance and has won several prestigious publication awards. At the University he has trained a number of teams that have won the CFA Institute Research Challenge in Switzerland as well as one team that became world champion.

"I think SFI has been pivotal for academia and the financial industry because it has allowed financial research to keep pace with the high quality of financial business being conducted in Switzerland, while fostering knowledge exchange among academics and practitioners. Such a platform creates great value for both theory and practice."