Our Perspective On Improving Efficiency:

Flexible Fintech Thinking Drives Accelerated U.S. Settlement Process

Today’s advances in the fintech space have enabled the financial industry to lower risk and cost by accelerating and streamlining key processes. For 45 years we have incorporated new technology and innovations to improve the financial marketplace. From our founding in the 1970s to help solve Wall Street’s paperwork crisis to the recent work in leading an industrywide effort to reduce the U.S. settlement cycle to T+2, a strong technology infrastructure has been at the core of our value to the industry.

Looking at ways to further improve the settlement system, we embraced transformative technology approaches to find a solution to modernize the settlement system without sacrificing speed or increasing costs.

The Depository Trust & Clearing Corporation (DTCC) has succeeded in balancing these objectives with an innovative, automated algorithmicdriven enhancement to the settlement cycle. This enhancement, known as night cycle reengineering, will double the current night cycle settlement rates, creating tremendous operational and processing efficiencies for clients while reducing risk for the entire industry.

After T+2

In a 2012 white paper, DTCC presented the case for several improvements to securities settlement processing, including shortening the cycle to T+2, that we and the industry implemented over the following few years. Earlier this year, we issued a follow-up white paper proposing further measures to accelerate settlement through optimization. But, unlike T+2, we found that these enhancements can be carried out more expeditiously and with less investment of time and money for market participants through more flexible thinking around technology innovations.

First on this optimization agenda is reengineering the DTC* night cycle. The DTC night cycle is a processing period that attempts to settle more than 1 million DTC and NSCC* transactions during the overnight period prior to settlement day. However, due to rigid and prescriptive legacy processing rules and algorithms, only approximately 45% of transactions eligible to settle in the night cycle actually settle in the night cycle. The remaining transactions “recycle” into DTC’s “day cycle” where they often require manual intervention to settle.

By reengineering the night cycle, we can not only improve processing efficiency, we can reduce risk for our clients by settling transactions earlier on settlement day and improving settlement finality. Additionally, an optimized night cycle opens opportunities to enhance other settlement processes that will free up capital for market participants.

Redefining the Problem

Approaching this project required us to collectively take a fresh look at the night cycle process, which hadn’t fundamentally changed in more than 20 years. A critical moment was reframing the problem. Rather than fitting as many transactions as possible into the current rules, we said, let’s take a holistic view of all transactions in our settlement ecosystem and develop a process that actively looks to settle as many transactions as possible. This refocus was a game changer for us.

We proceeded to develop a new algorithm designed to meet that objective and it produced dramatic improvements in night cycle settlement rates. Using this new algorithm, we saw transaction settlement rates as high as 90%, double the standard rate.

In the interest of reaching the most optimal solution, we also invited leading fintech providers to propose other options, but none outperformed our results. DTCC excelled in this effort thanks to our deep expertise and experience in the equities market along with strong internal team work, inherent in an effective Agile approach.

Taking an Agile Approach

A perfect algorithm that would provide a precise solution was impossible to derive, because of the unimaginably high costs, computing power and time required. So our multi-functional IT team developed a heuristic algorithm, one that met required levels of precision while solving the problem more quickly.

Today’s night cycle sequentially processes transactions. To devise our new algorithm, we upended this approach by loading a database that establishes nodes to create relationships between all transactions to each other, to counterparties and to our securities master file. This database of combined relationships enables us to apply a scoring model to all transactions, identifying which transactions have the highest value (in terms of the greatest impact on other transactions and counterparties), and which have the lowest value—that is, if removed from the night cycle have the least impact on everyone else.

Our algorithm removed those low-value exceptions from the cycle, which averaged around 10% of the total, leaving the remainder for processing and doubling our night cycle settlement rate to 90% on average, a tremendous win for the industry.

Close collaboration among subject-matter experts across multiple areas of DTCC, on the business and technology sides, was critical to our success. The work has also been a case study in the Agile method. Project leads identified who they needed on the team, pulled them in quickly, and the group conducted daily stand-ups to deliver direct feedback to and from the business. The direction of the work pivoted constantly based on what was learned at each step. The work also leveraged the Cloud to spin up the environment for testing and implementation, which gave us a preliminary result within 90 days.

Similarly, our collaboration with clients kept our objectives aligned with industry needs.

Future Possibilities

Going forward, we expect to harness machine learning to continually enhance night-cycle processing because our algorithm is configurable in design, lending itself to fine-tuning and on-going enhancement over time. Our subject-matter experts will remain part of the process, analyzing data to look for unusual patterns and opportunities for improvement. Machine learning can then spin out the patterns to understand them, creating a feedback loop that will enable ongoing tweaking of the settlement algorithm.

We expect to apply the same techniques to additional settlement optimization measures and even to processes beyond settlement, keeping DTCC at the forefront of technology innovation in the coming years.

*NSCC (The National Securities Clearing Corporation) and DTC (The Depository Trust Company) are wholly-owned subsidiaries of DTCC.