DTCC has recently launched a far-reaching initiative to enhance and modernize the U.S. equity settlement system, with the goal to optimize and further accelerate settlement to reduce costs and risk in the industry.
With a major part of this effort focused not just on streamlining and changing processing rules, but on improving the entire, underlying core clearance and settlement engine, DTCC’s technology teams have been working on this problem in new, creative and innovative ways.
Related: Creating an Alternate Settlement Model with Settlement Optimization
“Settlement is one of our most powerful and critical processes that has been around for over 40 years,” said Lynn Bishop, DTCC Managing Director and Chief Technology Officer. “While we have continually improved and enhanced our settlement processes over the years, what is different this time is leveraging flexible ‘fintech thinking’ to solve a traditional problem – using deep analytics and machine learning.”
Today, DTCC processes over 100 million financial transactions every dayusing a pre-defined set of processing rules and algorithms, however this process often leads to less than optimal settlement results. Currently settlement rates for night cycle transactions are approximately 45%, which means the other 55% must be held until later in the day to settle.
To create a more dynamic and intelligent processing environment, DTCC began a project last fall to explore how to reengineer the nighttime cycle using heuristic algorithms that could actively ‘look ahead’ to process as many transactions as possible — and as early in the processing cycle as possible.
A heuristic algorithm is one that is designed to solve a problem in a faster and more efficient fashion than traditional methods. In this case, instead of assuming transactions would result in fails and exceptions, the technology teams wrote algorithms that could predict and weed out the exceptions.
“By drastically changing the way we approached the problem – asking ‘why’ to get to a better solution – we’re now consistently hitting night cycle settlement rates of almost 90%,” said David Buksbaum, Executive Director of Application Architecture. “Maybe the algorithm initially sacrifices optimality, precision or completeness, but at the same time, it also uses continual machine-learning and advanced mathematical analysis to get to the solution faster. Our data models are showing extremely promising results.”
“Settlement Optimization has provided an exciting opportunity for us to rethink an entire core process, bringing together talent and experts from across different areas of data analytics, settlement and cloud computing,” said Bishop. “This is an exciting time for collaboration and fintech innovation at DTCC, and we haven’t even scratched the surface.”