Recently the Financial Industry Regulatory Authority hit independent broker-dealer Cambridge Investment Research and Merrill Lynch with fines totaling $850,000 for failing to properly supervise employees who were involved in the sale of mutual funds, and not properly monitoring the exchanges between retail marketers and exchange-traded note traders.

Notably, although Merrill Lynch did have a flagging system in place, built around a general lexicon search, it reportedly didn’t have sufficient reviewing practices organized — including no process for escalating reviews of private-public side communications that contained potential material information, or for enforcing required measures for separating traders and marketers in the global wealth and global banking and markets divisions.

Building the Chinese wall

This highlights a major challenge when it comes to compliance — particularly when it comes to the public and private side employees in financial institutions. The problem comes down to the fact that financial services firms frequently receive and handle information that counts as confidential or “insider” information (also known as MNPI – or Material Non-Public Information).  Traders or other agents who possess material information that has not been made public are prohibited from sharing it with others who do not have a need to know, even when they may be working on different sides of the same overall organization.

To maintain this divide, firms must erect information barriers around proper control of the flow of non-public information from one department to another. This is frequently called a Chinese wall, or informational firewall, referring to a virtual barrier that’s in place to stop the exchange of information that could result in illegal or ethically dubious activities.

A brief history of informational firewalls

The concept of Chinese walls has existed since the 1929 stock market crash when Congress first seriously discussed regulatory barriers separating investment bankers and brokers. However, the need for such divides has greatly increased over the past couple of decades, following the enacting of the Gramm-Leach-Bliley Act of 1999 (GLBA). This law, which helped empower many of today’s biggest financial powerhouses, repealed previous regulations that stopped firms from carrying out combinations of investing, banking, and insurance services.

Failing to maintain this informational divide can lead to some devastating consequences for financial institutions. In 2003, the Securities and Exchange Commission (SEC), National Association of Securities Dealers (NASD), New York Stock Exchange (NYSE), and other regulators announced that they had agreed a massive $1.4 billion settlement with 10 Wall Street firms for failing to mitigate against these conflicts of interest. Two well-known analysts were fined and given lifetime bans from participating in the securities industry. Among the stipulated changes were strengthened commitments to separate divisions within banking, to carry out extra-stringent monitoring and more.

Have the right tools in place

Today’s firms must be diligent in their stance on information sharing, ensuring that this happens only where absolutely required and lawful. As the Merrill Lynch example shows, having protective measures in place when it comes to monitoring isn’t enough. Lexicons, referring to a simple keyword or phrase searches, are massively outdated tools that can cause more problems than they solve. Lexicons yield a massive number of false positives (FPs), inundating system operators with high numbers of erroneous flagged messages, making them virtually valueless.

At Shield, we know that the world of detection doesn’t stay still. It’s not enough to simply set up lexicon-based models and hope that they will catch any potentially violating behavior that’s thrown at them. With that in mind, we continually add to, modify, and otherwise improve the detection models we used to provide updated coverage regarding the latest risk areas, along with new products, areas of business, mandated lines, and comments from regulators.

By knowing the employee department, job, management line, and other relevant information, Shield constructs specialized models to surveil specifically the eComm interactions between different groups of employees like traders and marketers within its global wealth and global banking and global banking divisions. We combine new age lexicon technologies, along with the latest AI innovations, to provide the best quality surveillance system of eComms. This improves the relevance of alerts, cutting down FPs, while also detecting infringements that regular lexicon searches or AI-only systems will miss. The technology we have developed is able to detect any hints in conversations between traders — or any other personnel requiring oversight — to identify when a potential breach is taking place. We are even capable of building models able to analyze complex interactions when two people are speaking in two territories with different regulations. This is a major problem and challenge in global operations.

Regulations separating the two sides of financial institutions are only going to get more stringent as time goes on. Fines are being handed out with increasing regularity — and merely paying lip service to detection through the installation of outdated detection tools isn’t enough. Firms must ensure that not only do they have a strong informational firewall in place but that they also have the detection tools present in case someone on either side finds a way over it.