McDonough School of Business
digital screen showing globe and various icons as well as 0 and 1 coding
News Story

Making the Case: FINRA Envisions Ways To Better Its Business

As a private company charged by the government with regulating 624,000 brokers and billions of investments, the Financial Industry Regulatory Authority (FINRA) has its organizational hands full. In an effort to help—and to take advantage of their data, analytics, machine learning, and predictive modeling expertise—McDonough’s Custom Executive Education team worked with analytics professors Sudipta Dasmohapatra and Greg Lyon to build an eight-week immersive program in partnership with the authority as a way of exploring new models of operations, and how best to use data and analytics to solve real issues at the nonprofit. 

We caught up with three FINRA employees and participants in the program and asked about their capstone project for the course. 

Pramit Das (P’19), Senior Director, Advertising Regulation 

The Problem: Das’ group reviews about 65,000 to 70,000 communications a year submitted by member firms. They then apply any FINRA or SEC rules to those communications. Das and fellow FINRA employees had the idea to use AI to create a risk assessment tool that can look at communications filed with the department to try and identify higher risk communications using data from decisions that had been made in the past. That model, however, couldn’t identify why communications were problematic. 

The Solution: In the Georgetown program, Das and his team wanted to take the FINRA model a step further. They built a model with “sentence-level classification,” essentially a tool that will crunch communications coming in and identify the problematic sentences or language. The accuracy rate, Das points out, is still not where he needs it to be, so he and his team will continue to refine. 

Christelle Niamke, Senior Director, Operations, Procedures and Standards in Member Supervision 

The Problem: Manual risk assessments take time and are exacerbated by a lack of transparency across business units, as well as a systematic prioritization of reviews. 

The Solution: Niamke’s group included individuals from technology, market regulation, and operations. The goal was to integrate information from disparate FINRA systems (alerts, risk scores, investigations) to highlight the risk of a firm sooner and more broadly across the organization. The benefits include prioritized reviews of firms that warrant immediate focus, bringing more timely impactful cases to enforcement partners and the SEC. 

Christian Collard, Senior Principal Investigator, Member Supervision, Special Investigations Unit 

The Problem: Identifying retail-investor suitability and best-interest concerns, and proving actual violations by stockbrokers, is a complex and time-consuming task. Existing conventions (e.g. 20% cost to equity and 6% turnover) are dated and overly prescriptive. 

The Solution: Collard and his team worked to study and develop an academic basis for establishing risk metrics and defining suitability/best-interest parameters, which then informed traditional risk-monitoring tools, surveillance models, and exploratory data analysis tools. The flexibility in the model allows them to be more inclusive and accurate in their identifications. 

This story was originally featured in the Georgetown Business Spring 2023 Magazine.

Custom Programs
Georgetown Business Magazine