Combatting Social Media Market Manipulation
Table of Contents
On February 18, 2021, the U.S. Congress held the first hearing related to the GameStop trading abnormalities. During this hearing, U.S. Representative David Scott (D-GA) raised the issue of social media’s effect on capital markets trading, calling its influence “a serious threat to the financial system.” He asked Robinhood CEO Vladimir Tenev if the trading platform monitored online forums for market manipulation risks. The answer to this question—not just from Robinhood but from most marketplace operators—was essentially, “no.”
As Tenev indicated during the hearing, Robinhood (and most trading platforms) lack access to the data that would be necessary to tie social media posts to individual identities. On Reddit’s end, although CEO Steve Huffman emphasized the platform’s moderation practices and efforts to curb distribution of false information, the lack of controls in place relating to social media and market manipulation was evident throughout the hearing.
The representatives present, and the public at large, treated this revelation with indignation. But this reality is no surprise to financial market professionals—and especially not to those involved in market surveillance efforts.
Surveillance Technology and Market Oversight
Many technology tools now available are designed to process unstructured data feeds—including those from social media—into useful information for both trading and oversight purposes. For market oversight efforts in particular, this functionality is a major differentiator for platforms, offering both communications monitoring and surveillance features.
Technologies such as natural language processing (NLP), behavioral pattern recognition in electronic/audio communications and sentiment analysis are increasingly being deployed by surveillance teams, whether those tools are built in-house or provided by a third party. More and more, these systems are incorporating machine learning or artificial intelligence-based processing in generating their surveillance alerts.
Especially in recent years, internal technology teams and surveillance technology providers have devoted growing levels of development resources to support social media monitoring, communications monitoring and alert generation, as well as efforts to increase trade surveillance alert integration capabilities. Recent Coalition Greenwich research continues to demonstrate the high demand for these evolved processing tools. Our data tells us repeatedly that budget allocations and adoption plans are prevalent and will continue to grow across sell-side and buy-side firms, as well as marketplace operations.
Solutions for Social Media Surveillance
Regarding these in-demand features, the surveillance technology space is ripe with fresh offerings. Some well-known examples of holistic surveillance solutions with highly developed unstructured data-processing features that are well suited for social media surveillance include:
- NICE Actimize, a communications surveillance leader and an early proponent of NLP-based alerts
- SteelEye, a new entrant focused on AI-based processing
- Nasdaq, a traditional market surveillance leader focusing on incorporation of unstructured data-based investigatory tools
On the communications-focused side, Behavox, Catelas, Smarsh (through its Digital Reasoning acquisition), Soteria, Verint, and VoxSmart are well known for their focus on innovative communications monitoring and alert generation.
Moreover, the potential for further integration between communications and trading surveillance efforts is promising and evident in such recently announced partnerships with Eventus (VoxSmart), FIS (Verint) and RIMES (Soteria).
Lack of Data Inputs is a Significant Obstacle
The growing use of innovative unstructured data processing technology bodes well for improved risk mitigation of social-media-led market manipulation in the coming years. A major remaining obstacle, however, is the lack of available data feeds and holistic integration capabilities. To operate effectively, these processing tools depends on extensive model training, historical scenario analysis and several other requirements—all of them dependent on large amounts of input data.
Without the ability to intake usable source material, and enough of it, even the most innovative AI- and NLP-based technologies cannot be effective. Few platform providers in the social media realm currently provide the type of feeds or access that would be useful for surveillance alert purposes.
The congressional testimony related to social media monitoring given by Tenev, Huffman and others was accurate and illuminates an important gap between capabilities and implementation in the regtech world: The tools needed to cast a much wider net on sources of market manipulation are both available and effective if properly used.
However, the data inputs themselves remain lacking. Even in this highly sophisticated world of surveillance technology, the adage “garbage in, garbage out” holds true. But it also means that the possibilities for more comprehensive monitoring of social media and, thus, more effective protection of market integrity, are within reach.
