How to Detect Fraud in Cryptocurrency Markets? Learning From Ambiguity and Cues

How do you like cryptocurrencies? How do you like fraud? As we know from the press, especially news reports on convictions such as that of SamBankman-Fried, the two questions are closely related. The unregulated nature of cryptocurrencies means that actions that are blatantly illegal in regular securities markets have a weaker legal proscription in cryptocurrency markets, and some people take advantage of it. That creates a problem for those who want to keep such markets healthy and make use of the currencies.

How they deal with this problem was the question studied by Bryan Spencer and Claus Rerup in research published in Administrative ScienceQuarterly. They examined a crypto-investment community that was hit by a series of fraudulent promotions by a group of actors who coordinated with each other to execute classic “pump-and-dump” schemes of minor cryptocurrencies. The problem, of course, was that the world of cryptocurrency movements and the world of online news are both ambiguous, so detecting coordinated fraud is difficult.

A major reason the researchers could detect that fraud was happening and that the investors were – after a while – becoming aware of this was that the researchers had access to the database containing public and group chats about the cryptocurrencies, which included the conversations among the fraudsters that were kept hidden from the other crypto-investors.
​
So, what did the investors do? Over time, they learnt. But they followed a path of nontraditional learning based on learning by making inferences from interpreting cues. They would question the intent behind the appearance of innocent-looking information (such as false analyst reports) supporting a currency. They would look for similarity between new information releases and earlier fraud cases. They would infer the intent behind the release of information even when they did not know who was behind the information release. And once they had made enough interpretation and inference, they would have stories ready about the true nature of the unfolding events and would be able to act in response. They acted by more systematically monitoring releases of information and by routinizing public responses to information releases that looked suspicious, so that other investors would become aware of the price pumping and could avoid entering.

In all, this article is a very interesting report on how a community can act to protect itself in the face of ambiguous information and repeated fraud attempts. But also, it is a reminder of why markets are regulated and why financial markets are regulated especially strictly. After all, the lengthy penalty of Sam Bankman-Fried was based on wrongdoing not in the cryptocurrency operations, but in regular and regulated financial markets.
​
Spencer, Bryan and Claus Rerup. 2024. The Dynamics of Inferential Interpretation in Experiential Learning: Deciphering Hidden Goals from Ambiguous Experience. Administrative Science Quarterly, forthcoming.
​

---